feat(explorer): add welcome screen and fix root path Invalid path error

- Add WelcomeScreen shown on app load; SKE brand button navigates back
- Fix serve_spa: empty root path was hitting dot-guard returning 400
  Invalid path instead of index.html or a welcome JSON response

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

.claude/init

@@ -0,0 +1 @@
+# Initialization

.claude/skills/init

@@ -0,0 +1 @@
+# Intialization

.claude/skills/semantica/init

@@ -0,0 +1 @@
+# Initialization

.github/workflows/docs.yml

@@ -63,7 +63,7 @@ jobs:
         continue-on-error: true
 
       - name: Upload artifact
-        uses: actions/upload-pages-artifact@v3
+        uses: actions/upload-pages-artifact@v5
         with:
           path: ./site
 

.gitignore

@@ -111,5 +111,12 @@ sample_data/
 # Test Results
 test_results.txt
 
+# Frontend workspace artifacts
+semantica-explorer/
+node_modules/
+
 # Frontend build artifacts (generated by Vite — do not track in git)
 semantica/static/
+
+# Local graph explorer test datasets
+demo_out/

CHANGELOG.md

@@ -7,6 +7,57 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+- **Feature: Graph Workspace declutter + calmer structural exploration** (PR #483 by @ZohaibHassan16, follow-up by @KaifAhmad1):
+  - Added a calmer default presentation for dense graphs: reduced label pressure, stronger inactive-state muting, and tuned zoom-tier visibility to improve readability during overview and structure navigation.
+  - Added display-edge aggregation with raw-edge bundle metadata retention, enabling cleaner visuals while preserving drill-down context for selected edges.
+  - Added grouped community view and neighborhood collapse/expand controls for high-degree local structures in Graph Workspace and Neighborhood panel flows.
+  - Extended graph selection/runtime state with display-state metadata (`groupedViewAvailable`, visible/collapsed neighbor counts, aggregated edge descriptors) for plugin and panel introspection.
+  - Added regression coverage for `resolveDisplayGraph` behavior in `explorer/tests/graphSceneState.display.test.ts`:
+    - parallel-edge aggregation in full view
+    - collapse behavior preserving active-path neighbors
+    - grouped community-node/community-edge projection behavior
+  - Follow-up merge resolution synced the PR branch with `main` after Explorer path migration (`semantica-explorer` -> `explorer`) and preserved PR #483 behavior in conflicted Graph Workspace files.
+
+- **Fix: DeepSeekProvider now uses OpenAI SDK instead of unmaintained deepseek SDK** (closes #482, PR #482 by @liling, review fixes by @KaifAhmad1):
+  - **Root cause**: The `deepseek` PyPI package has no `deepseek.Client`, causing `AttributeError` on every `DeepSeekProvider` instantiation. The DeepSeek API is OpenAI-compatible, so the `openai` SDK is the correct client.
+  - **`_init_client` rewritten**: Replaced `import deepseek; deepseek.Client(api_key=...)` with `from openai import OpenAI; OpenAI(api_key=..., base_url=self.base_url)`, matching the pattern already used by `NovitaProvider`.
+  - **`self.base_url` added to `__init__`**: Set to `"https://api.deepseek.com/v1"` (with `/v1` suffix required by the OpenAI SDK for correct endpoint resolution). This was missing from the original PR, causing a second `AttributeError` at `_init_client` call time.
+  - **`generate_typed` `verbose_mode` fix**: `verbose_mode` was referenced before assignment inside the instructor path. Added assignment `verbose_mode = kwargs.get("verbose", False) or self.config.get("verbose", False)` at the correct scope.
+  - **`pyproject.toml` updated**: `llm-deepseek` extra now declares `openai>=1.0.0` instead of the defunct `deepseek>=0.1.0`.
+  - **Warning message updated**: `_init_client` ImportError warning now references the `openai` library and `llm-openai` extra.
+  - **Instructor path improved**: Since `self.client` is now an `OpenAI` instance, the `isinstance(self.client, OpenAI)` check in `generate_typed` passes correctly, avoiding a redundant second client construction.
+  - 19 new tests in `tests/semantic_extract/test_pr482_deepseek_openai.py` across five suites: `TestDeepSeekProviderInit` (8 — covers `base_url`, OpenAI instantiation, no `deepseek` import, ImportError handling, `is_available`), `TestDeepSeekProviderGenerate` (5 — `generate`, `generate_structured`, no-client error paths), `TestDeepSeekInstructorPath` (1 — `isinstance` check), `TestVerboseModeAssignment` (4 — no NameError, verbose kwarg, config verbose, no-print default), `TestDeepSeekGenerateTypedInstructorIntegration` (1 — end-to-end instructor path reuses existing client).
+
+- **Performance: Indexed search for large knowledge graphs** (closes #467, PR #481 by @ZohaibHassan16, review fixes by @KaifAhmad1):
+  - **Root cause**: The previous `GraphSession.search()` ran a full O(n) scan over all nodes per query, serializing every node's properties to JSON for string matching. On a 118 k-node graph warm queries took 24–471 ms; a session with 500 k nodes was effectively unusable.
+  - **New `semantica/explorer/search_index.py`**: Purpose-built in-memory inverted index with three lookup tiers — exact-term index (full normalized strings), token index (individual words), and prefix index (2–12 character prefixes of every token). A linear secondary-scan fallback (capped at 12 k nodes) handles queries that miss all three tiers. An LRU result cache (128 slots, `OrderedDict`) serves repeat queries at zero cost. Warm query times on the same 118 k-node graph: 24 ms → 0.004 ms (exact), 471 ms → 0.009 ms (ID lookup), 475 ms → 0.002 ms (no-match).
+  - **`IndexedNodeDocument`** frozen dataclass stores per-node primary text (ID, content, curated alias keys: `label`, `name`, `pref_label`, `aliases`, `synonyms`, `display_name`, etc.), secondary text (remaining properties), token set, prefix expansions, confidence, and tags. Primary text prioritizes human-readable fields; secondary text covers the full property bag up to a 48-fragment cap.
+  - **Scoring**: exact ID match → 140, exact term → 120, primary-text substring → 78 + length bonus, token hit → 18, prefix hit → 10, multi-token bonus → 4 per hit. Ties broken deterministically on `(score, exactness, token_hits, node_id)`.
+  - **Mutation sync**: `GraphSession.add_node()`, `add_nodes()`, `add_edges()` and `add_node()` update the index incrementally. `handle_graph_mutation()` integrates with the WebSocket mutation bridge for live updates during reasoning, enrichment, and remote graph changes. `rebuild_search_index()` performs a full O(n) rebuild when needed (session init, merge, reload). `enrich.py` routes node/edge additions through `session.add_node`/`session.add_edge` so reasoning-inferred nodes are indexed immediately.
+  - **Review fixes applied**: replaced `list.sort()` per upsert with `bisect.insort()` (O(log n) vs O(n log n)); replaced `list.remove()` with `bisect.bisect_left` + `pop()` (O(log n) vs O(n)); added `with self._lock` in `handle_graph_mutation()` to prevent index races from the WebSocket thread; removed unnecessary source/target upserts on `add_edge()` (edges don't affect node text); sorted tag values in `_cache_key()` so `["a","b"]` and `["b","a"]` share a cache entry.
+  - **Follow-up fix by @KaifAhmad1 and @ZohaibHassan16**: restored `_ordered_node_ids` maintenance during upserts so `secondary_scan` fallback works for terms that are only present in non-curated properties; added regression test coverage to lock this behavior.
+  - 3 new tests: `test_search_exact_and_prefix` (exact match + prefix match), `test_search_filters_and_cache_stability` (type + confidence filter, identical repeated requests), `test_search_sees_new_nodes_after_mutation` (node added via `session.add_node()` immediately visible in search).
+  - 1 additional regression test: `test_search_secondary_scan_fallback_matches_non_curated_properties` (verifies fallback matching when a query term appears only in non-curated properties).
+- **Fix: Provenance traversal now includes multi-hop upstream ancestors + edge direction classification** (closes #470, PR #480 by @Sameer6305, review fixes by @KaifAhmad1):
+  - **Bug — upstream ancestors silently excluded**: `_build_provenance()` built a directed `nx.DiGraph` and seeded first-hop neighbors correctly, but the final subgraph extraction called `nx.ego_graph(..., undirected=False)`. With directed traversal, ego-graph expansion only follows outgoing edges from the focus node, so any node that *points into* the focus node (i.e. an upstream ancestor at depth ≥ 2) was invisible. For the chain `Source → Intermediate → node_id`, `Intermediate` appeared at hop 1 but `Source` was silently dropped. Fixed by changing to `undirected=True` — the radius expansion now traverses both incoming and outgoing edges while the underlying `DiGraph` is preserved, so edge source/target semantics remain correct.
+  - **Enhancement — edge direction classification**: `ProvenanceEdge` gains a `direction: str` field. Each edge in the provenance subgraph is classified relative to the focus node: `"upstream"` when `target == node_id` (edge flows into the focus node), `"downstream"` when `source == node_id` (edge flows out), and `"lateral"` for all other edges between non-focus neighbors. This lets consumers distinguish ancestor provenance from descendant impact without re-traversing the graph.
+  - **Enhancement — grouped markdown report**: `_render_markdown()` now groups lineage edges under separate `## Upstream`, `## Downstream`, and `## Lateral` sections instead of a flat `## Lineage Edges` list. Empty sections are omitted. This improves readability of exported provenance reports.
+  - **Schema consolidation**: `ProvenanceNode`, `ProvenanceEdge`, and `ProvenanceResponse` moved from inline definitions in `routes/provenance.py` to the shared `semantica/explorer/schemas.py`, matching the convention used by all other Explorer routes. `ProvenanceNode.parent_id` is now `Optional[str] = None`.
+  - **Merge conflicts resolved**: Resolved all conflict markers in `provenance.py`, `app.py`, and `.gitignore`; restored the complete router import set in `app.py` (`graph`, `sparql`, `temporal`, `vocabulary`) that the conflict had dropped.
+  - 2 new tests in `tests/explorer/test_provenance_route.py`: `test_build_provenance_direction_classification_chain` (asserts `Source` and `Intermediate` both appear for `Source → Intermediate → node_id`; verifies `Intermediate → node_id` classified as `"upstream"`) and `test_render_markdown_groups_edges_by_direction` (asserts grouped section headings and correct edge lines in output).
+
+- **Fix: `OWLExporter._export_owl_turtle` invalid Turtle syntax and silent data-property omission** (closes #478 by @KaifAhmad1):
+  - **Bug 1 — Invalid Turtle syntax**: `_export_owl_turtle` unconditionally wrote `rdfs:label` with a closing period (`.`), then appended `rdfs:subClassOf`, `rdfs:domain`, and `rdfs:range` triples after the closed block. Any RDF parser would reject the output. Fixed by introducing `_ttl_block(subject_uri, rdf_type, predicates)` — all predicate-object pairs for a subject are accumulated first, then joined with ` ;\n    ` and terminated with a single ` .`, producing valid Turtle in all cases.
+  - **Bug 2 — Data properties silently dropped**: `_export_owl_turtle` had loops for `classes` and `object_properties` but no loop for `data_properties`, so all `owl:DatatypeProperty` declarations were silently omitted. Added the missing loop, mirroring the existing object-property loop.
+  - **String escaping**: User-provided strings (`name`, `description`, `comment`, version) were embedded directly into Turtle string literals without escaping. A class named `John"s Class` or a comment containing a backslash or newline produced unparseable output. Added `_escape_ttl_str()` static method (escapes `"`, `\`, `\n`, `\r`, `\t`) applied at every `rdfs:label`, `rdfs:comment`, and `owl:versionInfo` site.
+  - **Null-check consistency**: All optional field reads now use `x = prop.get("field"); if x:` uniformly — eliminates the mixed pattern of `.get()` guards followed by direct `[]` access.
+  - 43 tests added in `tests/export/test_owl_exporter.py` across five suites: `TestTurtleSyntaxValidity` (5), `TestDataPropertiesInTurtle` (8), `TestTurtleHeader` (4), `TestTurtleStringEscaping` (16), `TestNullFieldHandling` (7), plus `TestObjectPropertyListDomainRange` (2) and `TestEquivalentClass` (1).
+
+- **Enhancement: Node distance semantics in path responses** (closes #472 by @KaifAhmad1): `PathResponse` now surfaces two new first-class fields — `hop_count: int` (equal to `len(path) - 1`; `0` for self-paths) and `distance_band: str` — so callers no longer need to count hops or implement band classification themselves. Four bands are defined: `"direct"` (0–1 hops), `"near"` (2–3), `"mid-range"` (4–6), `"distant"` (7+). The classification function `classify_path_distance()` lives in `semantica/utils/helpers.py` as the single source of truth; both the Explorer route and the visualizer import from it. `KGVisualizer.visualize_network()` gains an optional `highlight_path: list[str]` parameter: when provided, path edges are rendered as a separate orange trace with opacity and stroke width scaled to the distance band (direct: 1.0 / 4 px → distant: 0.35 / 1.5 px), while non-path edges render at reduced opacity underneath. Edge direction is respected — only the forward pairs `(A, B)` along the path are matched; reverse back-edges in directed graphs are not incorrectly highlighted. A logger warning is emitted when any node ID in `highlight_path` has no layout position, surfacing silent no-op mismatches. Frontend `PathResponse` type in `GraphInspectorPanel.tsx` and `GraphWorkspaceShell.tsx` extended with `hop_count: number` and `distance_band: "direct" | "near" | "mid-range" | "distant"`. All changes are additive; no existing fields removed. 10 new tests: 2 API-level (`test_response_includes_hop_count_and_distance_band`, `test_one_hop_path_is_direct`) and 8 unit tests covering all four band boundaries (0, 1, 2, 3, 4, 6, 7, 20 hops).
+
+- **Enhancement: Bidirectional path finding in Knowledge Explorer** (closes #469 by @KaifAhmad1): Path queries in the Explorer were direction-sensitive — querying B→A when only the edge A→B existed always returned no result, because `PathFinder._get_neighbors()` called `graph.neighbors(node)` which on a `nx.DiGraph` yields only successors. Added a `directed: bool = True` parameter to `bfs_shortest_path()` and `dijkstra_shortest_path()`. When `directed=False` a lightweight undirected view is built via `graph.to_undirected()` for the traversal pass only; the original directed edges are preserved and returned in the response. A `_make_undirected_view()` helper encapsulates the conversion and falls back safely for non-NetworkX graph types. The `/api/graph/node/{id}/path` route exposes the parameter as a query string flag (`?directed=false`); `PathResponse` gains a `directed: bool` field that echoes the mode used. Default is `True`, so all existing callers are unaffected. The route also gained an empty-path 404 guard — previously a traversal that found no path returned `200` with `path: []` instead of `404`. 21 new tests: 12 unit tests in `TestBidirectionalPathFinding` (`tests/kg/test_path_finder.py`) and 9 API-level tests in `TestBidirectionalPathRoute` (`tests/explorer/test_explorer_api.py`).
+
+- **Enhancement: Native `KnowledgeGraph` type support in `KGVisualizer`** (PR `kg` by @KaifAhmad1, closes #471): Added `semantica/kg/knowledge_graph.py` — a formal `KnowledgeGraph` dataclass (`entities`, `relationships`, `metadata`) that is now the canonical in-memory type produced and consumed by the Semantica KG pipeline. Exported from `semantica.kg`. `KGVisualizer` gains `_convert_knowledge_graph()` — an authoritative, non-mutating conversion path from `KnowledgeGraph` to the internal dict format — and `_normalize_graph()` now routes `isinstance(graph, KnowledgeGraph)` through it as an explicit fast-path before duck-typing. All five public entry points (`visualize_network`, `visualize_communities`, `visualize_centrality`, `visualize_entity_types`, `visualize_relationship_matrix`) accept `KnowledgeGraph` directly; no manual conversion required. All existing callers passing dicts or duck-typed objects are unaffected. 15 new tests in `TestFormalKnowledgeGraphType` (conversion shape, non-mutation, determinism, routing, all five entry points, import availability).
 - **Fix: `KGVisualizer` now accepts `KnowledgeGraph` objects in all `visualize_*` methods** (PR `visualization` by @KaifAhmad1, closes #458): All five public methods (`visualize_network`, `visualize_communities`, `visualize_centrality`, `visualize_entity_types`, `visualize_relationship_matrix`) previously called `graph.get("entities", [])`, silently producing no output when passed a non-dict object. Added `_normalize_graph()` which duck-types the input — dicts pass through unchanged; any object exposing `.entities` / `.relationships` attributes (e.g. the result of `GraphBuilder.build()`) is converted to the canonical dict form; anything else raises a clear `ProcessingError` naming the offending type. 21 tests added in `tests/visualization/test_kg_visualizer_normalize_graph.py`.
 
 - **Security: 12 vulnerability fixes across CRITICAL → LOW severity** (PR `security-enhancement` by @KaifAhmad1):

Dockerfile

@@ -1,4 +1,4 @@
-FROM node:20-alpine AS frontend-builder
+FROM node:25-alpine AS frontend-builder
 
 WORKDIR /app/semantica-explorer
 
@@ -13,7 +13,7 @@ COPY semantica-explorer/ ./
 RUN npm run build
 
 
-FROM python:3.12-slim AS runtime
+FROM python:3.14-slim AS runtime
 
 WORKDIR /app
 

README.md

@@ -46,7 +46,7 @@ Semantica is the **context and intelligence layer** you add on top of your exist
 - ✅ **Reasoning Engines** — forward chaining, Rete networks, deductive, abductive, and SPARQL. Explainable paths, not black boxes.
 - ✅ **Quality & Deduplication** — conflict detection, entity resolution, and pipeline validation built in.
 
-> Works alongside LangChain, LlamaIndex, AutoGen, CrewAI, and any LLM — Semantica is the **accountability layer** on top, not a replacement.
+> Works alongside **Agno** and any LLM — Semantica is the **accountability layer** on top, not a replacement. LangChain, LangGraph, CrewAI, and more coming soon.
 
 ```bash
 pip install semantica
@@ -170,6 +170,75 @@ Semantica ships **native plugin bundles** for Claude Code, Cursor, and Codex, an
 
 </table>
 
+### Agentic Frameworks
+
+Semantica integrates with **Agno** today. Coming soon: LangChain, LangGraph, CrewAI, LlamaIndex, AutoGen, OpenAI Agents SDK, Google ADK, and more.
+
+<table>
+<tr>
+<th colspan="8" align="left">✅ Supported</th>
+</tr>
+<tr>
+<td align="center" width="12.5%">
+<a href="https://github.com/agno-agi/agno"><img src="https://github.com/agno-agi.png?size=120" alt="Agno" width="48" height="48" /></a><br/>
+<strong>Agno</strong><br/>
+<sub>First-class · <code>pip install semantica[agno]</code></sub>
+</td>
+</tr>
+<tr>
+<th colspan="8" align="left">🔜 Coming Soon</th>
+</tr>
+<tr>
+<td align="center" width="12.5%">
+<a href="https://github.com/langchain-ai/langchain"><img src="https://github.com/langchain-ai.png?size=120" alt="LangChain" width="48" height="48" /></a><br/>
+<strong>LangChain</strong><br/>
+<sub>Coming soon</sub>
+</td>
+<td align="center" width="12.5%">
+<a href="https://github.com/langchain-ai/langgraph"><img src="https://github.com/langchain-ai.png?size=120" alt="LangGraph" width="48" height="48" /></a><br/>
+<strong>LangGraph</strong><br/>
+<sub>Coming soon</sub>
+</td>
+<td align="center" width="12.5%">
+<a href="https://github.com/crewAIInc/crewAI"><img src="https://github.com/crewAIInc.png?size=120" alt="CrewAI" width="48" height="48" /></a><br/>
+<strong>CrewAI</strong><br/>
+<sub>Coming soon</sub>
+</td>
+<td align="center" width="12.5%">
+<a href="https://github.com/run-llama/llama_index"><img src="https://github.com/run-llama.png?size=120" alt="LlamaIndex" width="48" height="48" /></a><br/>
+<strong>LlamaIndex</strong><br/>
+<sub>Coming soon</sub>
+</td>
+<td align="center" width="12.5%">
+<a href="https://github.com/microsoft/autogen"><img src="https://github.com/microsoft.png?size=120" alt="AutoGen" width="48" height="48" /></a><br/>
+<strong>AutoGen</strong><br/>
+<sub>Coming soon</sub>
+</td>
+<td align="center" width="12.5%">
+<a href="https://github.com/openai/openai-agents-python"><img src="https://github.com/openai.png?size=120" alt="OpenAI Agents SDK" width="48" height="48" /></a><br/>
+<strong>OpenAI Agents</strong><br/>
+<sub>Coming soon</sub>
+</td>
+<td align="center" width="12.5%">
+<a href="https://github.com/google/adk-python"><img src="https://github.com/google.png?size=120" alt="Google ADK" width="48" height="48" /></a><br/>
+<strong>Google ADK</strong><br/>
+<sub>Coming soon</sub>
+</td>
+</tr>
+</table>
+
+> **Agno — First-Class Integration** · `pip install semantica[agno]`
+>
+> Five integration modules live in [`integrations/agno/`](integrations/agno/):
+>
+> | Module | Class | What it does |
+> |---|---|---|
+> | `context_store.py` | `AgnoContextStore` | Graph-backed agent memory — store and retrieve structured context |
+> | `knowledge_graph.py` | `AgnoKnowledgeGraph` | Implements Agno's `AgentKnowledge` protocol; full extraction pipeline |
+> | `decision_kit.py` | `AgnoDecisionKit` | 6 decision-intelligence tools for Agno agents |
+> | `kg_toolkit.py` | `AgnoKGToolkit` | 7 KG pipeline tools (build, query, enrich, export) |
+> | `shared_context.py` | `AgnoSharedContext` | Shared context graph for multi-agent team coordination |
+
 ### Plugin Bundles (Claude Code · Cursor · Codex)
 
 Native plugin bundles live under [`plugins/`](plugins/). Each directory contains a `plugin.json`, `marketplace.json`, and `README.md`.
@@ -875,53 +944,6 @@ if result.valid:
 - **`semantica.llms`** — Groq, OpenAI, Novita AI, HuggingFace, LiteLLM
 - **[`explorer/`](explorer/)** — **Semantica Knowledge Explorer** — browser UI for live graph inspection, decisions, entity resolution, and ontology browsing (`npm run dev` in `explorer/`)
 
-
----
-
-## 🔌 Integrations
-
-### AI Coding Tools & IDEs
-
-Start the Semantica server (`python -m semantica.server`, port 8000) and point any tool at `http://localhost:8000`. Tools marked **Native plugin** also get 17 skills, 3 agents, and hook config out of the box.
-
-| Tool | Connection | Notes |
-|---|---|---|
-| [Claude Code](https://claude.com/product/claude-code) | **Native plugin** | `plugins/.claude-plugin/` — 17 skills, 3 agents, `hooks.json` |
-| [Cursor](https://cursor.com) | **Native plugin** | `plugins/.cursor-plugin/` — same 17 skills + 3 agents |
-| [Codex CLI](https://github.com/openai/codex) | **Native plugin** | `plugins/.codex-plugin/` — same 17 skills + 3 agents |
-| [Windsurf](https://windsurf.com) | **MCP server** + plugin | `plugins/.windsurf-plugin/` · add `python -m semantica.mcp_server` to `~/.codeium/windsurf/mcp_config.json` |
-| [Claude Desktop](https://claude.ai/download) | **MCP server** | Add `python -m semantica.mcp_server` to `claude_desktop_config.json` |
-| [VS Code](https://github.com/microsoft/vscode) | **MCP server** + plugin | `plugins/.vscode-plugin/` · add to `settings.json` under `mcp.servers` |
-| [GitHub Copilot](https://github.com/features/copilot) | REST API | Use via Copilot Chat custom tools |
-| [Cline](https://github.com/cline/cline) | **MCP server** + plugin | `plugins/.cline-plugin/` · add server in Cline MCP settings panel |
-| [Roo Code](https://github.com/RooCodeInc/Roo-Code) | **MCP server** | Add `python -m semantica.mcp_server` in Roo Code MCP settings |
-| [Continue](https://github.com/continuedev/continue) | **MCP server** + plugin | `plugins/.continue-plugin/` · add to `~/.continue/config.json` under `mcpServers` |
-| [Goose](https://github.com/block/goose) | REST API | Add to Goose toolset config |
-| [Kilo Code](https://github.com/Kilo-Org/kilocode) | REST API | Add as custom REST tool |
-| [Aider](https://github.com/Aider-AI/aider) | REST API | Pass context from the API into prompts |
-| [Amazon Q Developer](https://github.com/aws/amazon-q-developer-cli) | REST API | Use via Q Developer custom tools |
-| [Zed](https://zed.dev) | REST API | Integrate via Zed assistant context |
-| Any agent | REST API | 109 endpoints — drop-in with any HTTP client |
-
-### REST API Server
-
-Run `python -m semantica.server` (or `python -m semantica`) — FastAPI on port 8000 with the following route groups:
-
-| Route group | Module | Endpoints |
-|---|---|---|
-| `/api/graph` | `routes/graph.py` | Nodes, edges, traversal, graph topology |
-| `/api/analytics` | `routes/analytics.py` | Centrality, communities, metrics |
-| `/api/decisions` | `routes/decisions.py` | Decision CRUD, precedent search, causal chains |
-| `/api/temporal` | `routes/temporal.py` | Point-in-time queries, snapshots, timelines |
-| `/api/export` | `routes/export_import.py` | Import/export in RDF, Parquet, JSON, CSV, GraphML |
-| `/api/annotations` | `routes/annotations.py` | Entity and edge annotation |
-| `/api/enrich` | `routes/enrich.py` | Graph enrichment — embeddings, vectors, metadata |
-| `/api/sparql` | `routes/sparql.py` | SPARQL query execution |
-| `/api/provenance` | `routes/provenance.py` | Data lineage and audit trails |
-| `/api/vocabulary` | `routes/vocabulary.py` | Ontology, SKOS concepts, schema definitions |
-| `/ws` | `ws.py` | WebSocket — real-time graph mutation events |
-| `/health` | `server.py` | Health check |
-
 ### Graph Databases
 - **Neo4j** — Cypher queries via `semantica.graph_store`
 - **FalkorDB** — native support; `DecisionQuery` and `CausalChainAnalyzer` work directly with FalkorDB row/header shapes
@@ -952,67 +974,6 @@ Run `python -m semantica.server` (or `python -m semantica`) — FastAPI on port
 - **HuggingFace** — local and hosted models via `HuggingFaceProvider`
 - **Ollama** — local models including remote server support
 
-### Agentic Frameworks
-
-Semantica complements — not replaces — every major agentic framework. Use it as the accountability layer on top.
-
-<table>
-<tr>
-<td align="center" width="12.5%">
-<a href="https://github.com/agno-agi/agno"><img src="https://github.com/agno-agi.png?size=120" alt="Agno" width="40" height="40" /></a><br/>
-<strong>Agno</strong><br/>
-<sub>First-class · <code>pip install semantica[agno]</code></sub>
-</td>
-<td align="center" width="12.5%">
-<a href="https://github.com/langchain-ai/langchain"><img src="https://github.com/langchain-ai.png?size=120" alt="LangChain" width="40" height="40" /></a><br/>
-<strong>LangChain</strong><br/>
-<sub>Context layer</sub>
-</td>
-<td align="center" width="12.5%">
-<a href="https://github.com/langchain-ai/langgraph"><img src="https://github.com/langchain-ai.png?size=120" alt="LangGraph" width="40" height="40" /></a><br/>
-<strong>LangGraph</strong><br/>
-<sub>Stateful agent graph</sub>
-</td>
-<td align="center" width="12.5%">
-<a href="https://github.com/run-llama/llama_index"><img src="https://github.com/run-llama.png?size=120" alt="LlamaIndex" width="40" height="40" /></a><br/>
-<strong>LlamaIndex</strong><br/>
-<sub>GraphRAG retriever</sub>
-</td>
-<td align="center" width="12.5%">
-<a href="https://github.com/microsoft/autogen"><img src="https://github.com/microsoft.png?size=120" alt="AutoGen" width="40" height="40" /></a><br/>
-<strong>AutoGen</strong><br/>
-<sub>Shared context graph</sub>
-</td>
-<td align="center" width="12.5%">
-<a href="https://github.com/crewAIInc/crewAI"><img src="https://github.com/crewAIInc.png?size=120" alt="CrewAI" width="40" height="40" /></a><br/>
-<strong>CrewAI</strong><br/>
-<sub>Decision + provenance</sub>
-</td>
-<td align="center" width="12.5%">
-<a href="https://github.com/openai/openai-agents-python"><img src="https://github.com/openai.png?size=120" alt="OpenAI Agents SDK" width="40" height="40" /></a><br/>
-<strong>OpenAI Agents</strong><br/>
-<sub>Context + KG tools</sub>
-</td>
-<td align="center" width="12.5%">
-<a href="https://github.com/google/adk-python"><img src="https://github.com/google.png?size=120" alt="Google ADK" width="40" height="40" /></a><br/>
-<strong>Google ADK</strong><br/>
-<sub>Context layer</sub>
-</td>
-</tr>
-</table>
-
-> **Agno — First-Class Integration** · `pip install semantica[agno]`
->
-> Five integration modules live in [`integrations/agno/`](integrations/agno/):
->
-> | Module | Class | What it does |
-> |---|---|---|
-> | `context_store.py` | `AgnoContextStore` | Graph-backed agent memory — store and retrieve structured context |
-> | `knowledge_graph.py` | `AgnoKnowledgeGraph` | Implements Agno's `AgentKnowledge` protocol; full extraction pipeline |
-> | `decision_kit.py` | `AgnoDecisionKit` | 6 decision-intelligence tools for Agno agents |
-> | `kg_toolkit.py` | `AgnoKGToolkit` | 7 KG pipeline tools (build, query, enrich, export) |
-> | `shared_context.py` | `AgnoSharedContext` | Shared context graph for multi-agent team coordination |
-
 ---
 
 ## 🛠️ Installation

RELEASE_NOTES.md

@@ -1,282 +0,0 @@
-# Semantica v0.3.0 — Release Notes
-
-**Released:** 2026-03-10
-**PyPI:** `pip install semantica`
-**Tag:** [v0.3.0](https://github.com/Hawksight-AI/semantica/releases/tag/v0.3.0)
-**Classification:** Production/Stable
-
-> First stable, full public release of Semantica. Covers everything shipped across three release stages: 0.3.0-alpha (2026-02-19), 0.3.0-beta (2026-03-07), and 0.3.0 stable (2026-03-10).
-
----
-
-## Contributors
-
-| Contributor | Role |
-|------------|------|
-| [@KaifAhmad1](https://github.com/KaifAhmad1) | Lead maintainer — context graph, decision intelligence, KG algorithms, semantic extraction, pipeline, provenance, bug fixes, release management |
-| [@ZohaibHassan16](https://github.com/ZohaibHassan16) | Deduplication v2 suite (candidate generation, two-stage scoring, semantic dedup), incremental/delta processing, benchmark suite |
-| [@Sameer6305](https://github.com/Sameer6305) | Apache AGE backend, PgVector store, Snowflake connector, Apache Arrow export |
-| [@tibisabau](https://github.com/tibisabau) | ArangoDB AQL export, Apache Parquet export |
-| [@d4ndr4d3](https://github.com/d4ndr4d3) | ResourceScheduler deadlock fix |
-
----
-
-## v0.3.0 — Stable (2026-03-10)
-
-### Context Graph Feature Completeness
-
-**Temporal Validity Windows** (by @KaifAhmad1)
-
-Nodes and edges now carry first-class `valid_from` / `valid_until` ISO datetime fields. These are stored directly on `ContextNode` and `ContextEdge` dataclasses — not in metadata — and survive full serialisation round-trips through `save_to_file()` / `load_from_file()` and `to_dict()` / `from_dict()`.
-
-- `ContextNode.is_active(at_time=None)` and `ContextEdge.is_active(at_time=None)` — returns `True` if the node/edge is live at the given time (defaults to now). Handles both tz-aware and tz-naive datetime inputs correctly.
-- `ContextGraph.find_active_nodes(node_type=None, at_time=None)` — filters the entire graph and returns only nodes within their validity window.
-- `add_node(valid_from=..., valid_until=...)` and `add_edge(valid_from=..., valid_until=...)` — pass validity fields directly in the call signature.
-- Bug fix: `is_active()` previously crashed with `TypeError` when passed a tz-aware `datetime` (e.g. `datetime.now(timezone.utc)`). Fixed by normalising all inputs to tz-naive UTC via a new `_parse_iso_dt()` helper.
-- Bug fix: validity fields were silently lost in `add_nodes()`, `add_edges()`, `to_dict()`, and `from_dict()`. All four paths now correctly preserve and restore them.
-
-**Weighted Multi-Hop BFS** (by @KaifAhmad1)
-
-`ContextGraph.get_neighbors(node_id, hops=1, relationship_types=None, min_weight=0.0)` now accepts a `min_weight` threshold. Any edge with weight below the threshold is skipped during BFS traversal, allowing callers to confine multi-hop queries to high-confidence causal links. Default `0.0` is fully backward-compatible.
-
-**Cross-Graph Navigation** (by @KaifAhmad1)
-
-Separate `ContextGraph` instances can now be linked and navigated between — hierarchically, like separate knowledge domains that reference each other.
-
-- `link_graph(other_graph, source_node_id, target_node_id, link_type="CROSS_GRAPH") -> str` — creates a navigable bridge and returns a `link_id`. Records a dedicated `"cross_graph_link"` typed marker node internally (not a phantom `"entity"`) and a marker edge.
-- `navigate_to(link_id) -> (other_graph, target_node_id)` — jumps to the target graph and entry node for a given link.
-- `graph_id` field — each `ContextGraph` now carries a stable UUID so instances can identify each other across save/load.
-- `save_to_file()` — now writes a `links` section alongside nodes and edges, containing `link_id`, `source_node_id`, `target_node_id`, and `other_graph_id` for every cross-graph link.
-- `load_from_file()` — restores `graph_id` and populates `_unresolved_links` from the `links` section.
-- `resolve_links(registry: Dict[str, ContextGraph]) -> int` — reconnects unresolved links post-load. Pass `{graph_id: graph_instance}` for each linked graph; returns the count of successfully resolved links. `navigate_to()` raises a clear `KeyError` with a `resolve_links()` hint if called before resolution.
-- Bug fix: the previous implementation auto-created the synthetic marker target as an `"entity"` node (phantom pollution). Fixed by explicitly pre-creating a `"cross_graph_link"` typed `ContextNode` before the marker edge.
-- 14 new tests in `tests/context/test_cross_graph_navigation.py` covering all scenarios including full save/load round-trips with partial registry resolution.
-
-**Other Fixes** (by @KaifAhmad1)
-
-- `PipelineBuilder.add_step()` return type annotation corrected from `"PipelineBuilder"` to `"PipelineStep"` — the implementation was already correct; only the annotation and docstring were stale.
-- `test_hybrid_search_performance` timing computation fixed — now accumulates a true `search_times` list instead of reusing the last loop iteration's `start_time`; threshold relaxed to `< 5.0s` for real `sentence-transformers` (384-dim) latency on development machines.
-
-**Test Coverage Added**
-
-- 14 cross-graph navigation tests (`tests/context/test_cross_graph_navigation.py`)
-- **Total: 335 context tests, 886+ tests across all modules — 0 failures**
-
----
-
-## v0.3.0-beta — Beta (2026-03-07)
-
-### Semantic Extraction Fixes
-
-**Multi-Founder LLM Extraction & Reasoner Inference Fix** (PR #354, by @KaifAhmad1)
-
-- `_parse_relation_result` in `methods.py` — unmatched subjects/objects now produce a synthetic `UNKNOWN` entity instead of silently dropping the relation. All co-founders returned by the LLM are preserved in the output.
-- Duplicate relation fix — an orphaned legacy block that appended every relation twice has been removed.
-- `extraction_method` parameter added — typed extraction paths now correctly record `"llm_typed"` in relation metadata instead of `"llm"`.
-- `_match_pattern` in `reasoner.py` rewritten — splits patterns on `?var` placeholders first, then escapes only literal segments. Pre-bound variables resolve to exact literals, repeated variables use backreferences, non-greedy `.+?` prevents over-consumption of separators.
-- Added `tests/reasoning/test_reasoner.py` (4 tests) and `tests/semantic_extract/test_relation_extractor.py` (6 tests).
-
-**TTL Export Alias Fix** (PR #355, by @KaifAhmad1)
-
-- `RDFExporter` now accepts `"ttl"`, `"nt"`, `"xml"`, `"rdf"`, and `"json-ld"` as format aliases in `export_to_rdf()`. Aliases resolve before format validation — zero public API changes.
-- Added `tests/export/test_rdf_exporter.py` (8 tests).
-
-### Incremental / Delta Processing
-
-**Native Delta Computation** (PR #349, by @ZohaibHassan16, reviewed and fixed by @KaifAhmad1)
-
-- Native SPARQL-based diff between graph snapshots — only changed triples flow through the pipeline.
-- `delta_mode` configuration in `PipelineBuilder` for near-real-time workloads.
-- Version snapshot management with graph URI tracking and metadata storage.
-- `prune_versions()` for automatic snapshot retention cleanup.
-- Bug fixes: corrected SPARQL variable order, fixed class references, resolved duplicate dictionary keys.
-
-### Deduplication v2
-
-**Candidate Generation v2** (PR #338, by @ZohaibHassan16)
-
-- New opt-in strategies: `blocking_v2` and `hybrid_v2`, replacing O(N²) pair enumeration.
-- Multi-key blocking with normalised token prefixes, type-aware keys, and optional phonetic (Soundex) blocking.
-- Deterministic `max_candidates_per_entity` budgeting with stable sorting.
-- **63.6% faster** in worst-case scenarios (0.259s → 0.094s for 100 entities).
-
-**Two-Stage Scoring Prefilter** (PR #339, by @ZohaibHassan16)
-
-- Fast gates for type mismatch, name-length ratio, and token overlap eliminate expensive semantic scoring for obvious non-matches.
-- Configurable thresholds: `min_length_ratio`, `min_token_overlap_ratio`, `required_shared_token`.
-- **18–25% faster** batch processing with prefilter enabled (`prefilter_enabled=False` by default).
-
-**Semantic Relationship Deduplication v2** (PR #340, by @ZohaibHassan16, fixes by @KaifAhmad1)
-
-- Canonicalisation engine with predicate synonym mapping (`works_for` → `employed_by`).
-- O(1) hash matching for exact canonical signatures.
-- Weighted scoring: 60% predicate + 40% object composition with explainable `semantic_match_score`.
-- **6.98x faster** than legacy mode (83ms vs 579ms).
-- `dedup_triplets()` infinite recursion bug fixed; function is now a first-class API in `methods.py`.
-
-**Deduplication v2 Migration Guide** (PR #344, by @ZohaibHassan16, fixes by @KaifAhmad1)
-
-- Comprehensive `MIGRATION_V2.md` documenting all v2 strategies with code examples.
-- Full backward compatibility maintained — legacy mode remains the default.
-
-### Export Formats
-
-**ArangoDB AQL Export** (PR #342, by @tibisabau)
-
-- Full AQL INSERT statement generation for vertices and edges.
-- Configurable collection names with validation and sanitisation; batch processing (default: 1000).
-- `export_arango()` convenience function; `.aql` auto-detection in the unified exporter.
-- 17 tests, 100% pass rate.
-
-**Apache Parquet Export** (PR #343, by @tibisabau)
-
-- Columnar storage format with configurable compression: snappy, gzip, brotli, zstd, lz4, none.
-- Explicit Apache Arrow schemas with type safety; field normalisation for varied naming conventions.
-- `export_parquet()` convenience function; `.parquet` auto-detection.
-- Analytics-ready for pandas, Spark, Snowflake, BigQuery, Databricks.
-- 25 tests, 100% pass rate.
-
-### Bug Fixes & Test Suite Stabilisation
-
-**Test Suite Fixes** (by @KaifAhmad1)
-
-Context module:
-- `retrieve_decision_precedents` — gated entity extraction on `use_hybrid_search=True` correctly.
-- `_extract_entities_from_query` — now uses `word[0].isupper()` to capture camelCase identifiers like `CreditCard`.
-- Added missing `expand_context` (BFS traversal) and `_get_decision_query` methods.
-- Fixed `hybrid_retrieval`, `dynamic_context_traversal`, and `multi_hop_context_assembly` for correct single-pass BFS.
-- Fixed `_retrieve_from_vector` fallback to `result["metadata"]["content"]` to prevent empty content and negative re-ranking scores.
-
-KG module:
-- `calculate_pagerank` — added `alpha`/`max_iter` aliases; return format changed to `{"centrality": scores, "rankings": sorted_list}`.
-- `community_detector._to_networkx` — now returns a NetworkX graph directly when one is passed (previously lost all edges).
-- Added 9 domain-specific tracking methods to `AlgorithmTrackerWithProvenance`.
-- Created `provenance_tracker.py` with `ProvenanceTracker` (`track_entity`, `get_all_sources`, `clear`).
-
-Pipeline module:
-- Retry loop fixed — now correctly iterates to `max_retries`.
-- `RecoveryAction` dataclass and `handle_failure(error, policy, retry_count)` added with LINEAR, EXPONENTIAL, and FIXED strategies.
-- `add_step()` fixed to return the created `PipelineStep`.
-- `validate` added as alias for `validate_pipeline` in `PipelineValidator`.
-
-Other:
-- Fixed `NameError` for missing `Type` import in `utils/helpers.py`.
-- Vector store performance threshold relaxed (100ms → 500ms per decision for development machines).
-- Windows cp1252 encoding fix in test files (emoji → ASCII).
-- `ProvenanceTracker` added to `semantica/kg/__init__.py` exports.
-
-**Results: ~840 tests passing, 36 skipped (external services), 0 failed**
-
----
-
-## v0.3.0-alpha — Alpha (2026-02-19)
-
-### Context & Decision Intelligence
-
-**Context Engineering Enhancement** (PR #307, by @KaifAhmad1)
-
-The foundational 0.3.0 feature — complete overhaul of the context module for production-grade decision intelligence:
-
-- Full decision lifecycle: `record_decision()` → `add_decision()` → `add_causal_relationship()` → `trace_decision_chain()` → `analyze_decision_impact()` → `analyze_decision_influence()` → `find_similar_decisions()`
-- `AgentContext` unified wrapper with granular feature flags: `decision_tracking`, `kg_algorithms`, `graph_expansion`; methods: `store()`, `retrieve()`, `get_conversation_history()`, `get_statistics()`, `capture_cross_system_inputs()`
-- `AgentMemory` with working, conversation, and long-term memory tiers
-- `PolicyEngine` with versioned policy nodes, compliance checking (`check_decision_rules()`), and `PolicyException` model
-- Hybrid precedent search combining vector, structural, and category similarity with configurable weights
-- Decision influence analysis via centrality measures and causal chain tracking
-- GraphStore validation preventing runtime failures; secure logging
-- 9 critical bug fixes across logging, security, audit trails, API compatibility, Cypher queries, centrality access, validation
-
-**Context Decision Tracking Fixes** (PR #315, by @KaifAhmad1)
-
-- Fixed empty/None decision ID handling in `add_decision()`
-- Fixed None metadata handling preventing `TypeError`
-- Fixed causal chain depth logic and node exclusion
-- Fixed nonexistent node handling in `add_causal_relationship()`
-- Fixed precedent search direction in `find_precedents()`
-- Added missing `properties` field in `to_dict()`; added `from_dict()` method
-- Fixed UUID generation across all decision models
-- All 71 context tests passing
-
-### Knowledge Graph Algorithms
-
-**Improved Graph Algorithms** (PR #292, by @KaifAhmad1)
-
-- 30+ graph algorithms across 7 categories
-- Node embeddings: Node2Vec, DeepWalk, Word2Vec via `NodeEmbedder`
-- Similarity: cosine, Euclidean, Manhattan, Correlation via `SimilarityCalculator`
-- Path finding: Dijkstra, A*, BFS, K-shortest paths via `PathFinder`
-- Link prediction: preferential attachment, Jaccard, Adamic-Adar via `LinkPredictor`
-- Centrality: degree, betweenness, closeness, PageRank via `CentralityAnalyzer`
-- Community detection: Louvain, Leiden, label propagation via `CommunityDetector`
-- Connectivity: components, bridges, density via `ConnectivityAnalyzer`
-- `GraphBuilderWithProvenance` and `AlgorithmTrackerWithProvenance` with full execution metadata
-
-**Improved Vector Store for Decision Tracking** (PR #293, by @KaifAhmad1)
-
-- `DecisionEmbeddingPipeline` with semantic and structural embeddings
-- `HybridSimilarityCalculator` with configurable weights (semantic: 0.7, structural: 0.3)
-- `ContextRetriever` with multi-hop reasoning
-- Convenience API: `quick_decision()`, `find_precedents()`, `explain()`, `similar_to()`, `batch_decisions()`, `filter_decisions()`
-- 34+ tests; performance: 0.028s per decision, 0.031s search, ~0.8KB memory per decision
-
-### Graph Database Backends
-
-**Apache AGE Backend Security Fixes** (PR #311, by @Sameer6305, fixes by @KaifAhmad1)
-
-- `AgeStore` class with `GraphStore` API compatibility (openCypher via SQL on PostgreSQL)
-- SQL injection vulnerabilities fixed with input validation
-- psycopg2-binary dependency and migration guide added
-- Fixed parameter replacement and test mock leakage
-
-**PgVector Store Support** (PR #303, by @Sameer6305, @KaifAhmad1)
-
-- Native PostgreSQL vector storage using the pgvector extension
-- Multiple distance metrics: cosine, L2/Euclidean, inner product
-- HNSW and IVFFlat indexing for approximate nearest neighbour search
-- JSONB metadata storage with flexible filtering; batch operations
-- Connection pooling with psycopg3/psycopg2 fallback
-- SQL injection protection via `psycopg_sql.SQL()`; idempotent index and table management
-- 36+ tests with Docker integration
-
-### Infrastructure
-
-**ResourceScheduler Deadlock Fix** (PR #299, #301, by @d4ndr4d3, @KaifAhmad1)
-
-- Replaced `threading.Lock()` with `threading.RLock()` to fix nested lock acquisition deadlock in `allocate_resources()`
-- Added `ValidationError` when no resources can be allocated
-- Progress tracking updates moved outside lock scope
-- 6 regression tests for deadlock prevention
-
-**Security Configuration** (by @KaifAhmad1)
-
-- Dependabot bi-weekly security updates with manual review
-- Automated security scans (Bandit, Safety, Semgrep) on schedule
-- Security-critical package grouping; zero auto-merge policy
-
----
-
-## Summary by the Numbers
-
-| Metric | Value |
-|--------|-------|
-| Total tests passing | **886+** |
-| Test failures | **0** |
-| Context tests | 335 |
-| KG tests | ~430 |
-| Semantic extraction tests | 70 (9 skipped — external LLM APIs) |
-| Reasoning tests | 19 |
-| Real-world scenario tests | 85 |
-| PyPI classifier | Production/Stable |
-| Python support | 3.8 – 3.12 |
-
----
-
-## Upgrade
-
-```bash
-pip install --upgrade semantica
-```
-
-No breaking changes. All new parameters have safe defaults and all new methods are additive.
-
-See [CHANGELOG.md](CHANGELOG.md) for the full line-by-line diff.

STRATEGIES_SUMMARY.md

@@ -1,105 +0,0 @@
-# Deduplication & Conflict Resolution Strategies Summary
-
-## Quick Reference by Use Case
-
-| Use Case | Deduplication Method | Merge Strategy | Conflict Detection | Conflict Resolution |
-|----------|---------------------|----------------|-------------------|---------------------|
-| **Finance** |
-| `01_Financial_Data_Integration_MCP` | `DuplicateDetector` (incremental) | `keep_highest_confidence` | `temporal` | `most_recent` |
-| `02_Fraud_Detection` | `ClusterBuilder` (graph_based) | `merge_all` | `logical` | `expert_review` |
-| **Biomedical** |
-| `01_Drug_Discovery_Pipeline` | `EntityResolver` (semantic) | - | `relationship` | `voting` |
-| `02_Genomic_Variant_Analysis` | `DuplicateDetector` (group) | `keep_most_complete` | `value` | `credibility_weighted` |
-| **Cybersecurity** |
-| `01_Real_Time_Anomaly_Detection` | `DuplicateDetector` (pairwise) | `keep_first` | `entity` | `first_seen` |
-| `02_Threat_Intelligence_Hybrid_RAG` | `EntityResolver` (exact) | - | `type` | `highest_confidence` |
-| **Blockchain** |
-| `01_DeFi_Protocol_Intelligence` | `DuplicateDetector` (group) | `keep_last` | `relationship` | `voting` |
-| `02_Transaction_Network_Analysis` | `ClusterBuilder` (hierarchical) | `keep_most_complete` | `temporal` | `most_recent` |
-| **Intelligence** |
-| `01_Criminal_Network_Analysis` | `EntityResolver` (fuzzy) | - | `value` | `credibility_weighted` |
-| `02_Intelligence_Analysis_Orchestrator_Worker` | `DuplicateDetector` (batch) | `merge_all` | `entity` | `voting` |
-| **Renewable Energy** |
-| `01_Energy_Market_Analysis` | `DuplicateDetector` (pairwise) | `keep_highest_confidence` | `temporal` | `most_recent` |
-| **Supply Chain** |
-| `01_Supply_Chain_Data_Integration` | `DuplicateDetector` (incremental) | `keep_most_complete` | `value` | `credibility_weighted` |
-
----
-
-## Strategy Rationale by Domain
-
-### Finance
-- **Financial Data Integration**: Incremental for streaming data; most_recent for time-sensitive financial data
-- **Fraud Detection**: Graph-based clustering for fraud groups; expert_review for fraud assessment
-
-### Biomedical
-- **Drug Discovery**: Semantic matching for drug compounds; voting for research source aggregation
-- **Genomic Variants**: Group method for related variants; credibility weighting for research sources
-
-### Cybersecurity
-- **Real-Time Anomaly**: Pairwise for real-time streams; keep_first for first detection priority
-- **Threat Intelligence**: Exact matching for IOCs; highest_confidence for threat classification
-
-### Blockchain
-- **DeFi Protocols**: Group method for related protocols; keep_last for latest protocol info
-- **Transaction Networks**: Hierarchical clustering for nested groups; temporal for time-sensitive data
-
-### Intelligence
-- **Criminal Networks**: Fuzzy matching for intelligence data; credibility weighting for intelligence sources
-- **Intelligence Analysis**: Batch for multi-source integration; merge_all to combine all intelligence sources
-
-### Renewable Energy
-- **Energy Markets**: Pairwise for real-time market data; most_recent for time-sensitive energy data
-
-### Supply Chain
-- **Supply Chain Integration**: Incremental for continuous updates; credibility weighting for supply chain sources
-
----
-
-## Method Distribution
-
-### Deduplication Methods (9 total)
-- `pairwise`: 2 notebooks (real-time processing)
-- `batch`: 3 notebooks (large datasets)
-- `incremental`: 2 notebooks (streaming/continuous)
-- `group`: 2 notebooks (related entities)
-- `graph_based` (ClusterBuilder): 2 notebooks (interconnected entities)
-- `hierarchical` (ClusterBuilder): 1 notebook (nested groups)
-- `exact` (EntityResolver): 1 notebook (exact matching)
-- `semantic` (EntityResolver): 2 notebooks (semantic similarity)
-- `fuzzy` (EntityResolver): 1 notebook (fuzzy matching)
-
-### Merge Strategies (5 total)
-- `keep_first`: 1 notebook (first detection priority)
-- `keep_last`: 1 notebook (latest information)
-- `keep_most_complete`: 5 notebooks (preserve all details)
-- `keep_highest_confidence`: 2 notebooks (most reliable data)
-- `merge_all`: 3 notebooks (combine all information)
-
-### Conflict Detection Methods (6 total)
-- `value`: 4 notebooks (property value conflicts)
-- `type`: 2 notebooks (type/classification conflicts)
-- `entity`: 2 notebooks (entity-wide conflicts)
-- `relationship`: 3 notebooks (relationship conflicts)
-- `temporal`: 3 notebooks (time-sensitive conflicts)
-- `logical`: 2 notebooks (logical inconsistencies)
-
-### Conflict Resolution Strategies (6 total)
-- `voting`: 5 notebooks (majority vote)
-- `credibility_weighted`: 4 notebooks (source credibility)
-- `most_recent`: 3 notebooks (latest data)
-- `first_seen`: 1 notebook (first detection)
-- `highest_confidence`: 2 notebooks (most confident)
-- `expert_review`: 1 notebook (manual review)
-
----
-
-## Key Patterns
-
-1. **Real-Time Systems**: Use `pairwise` + `keep_first` + `first_seen`
-2. **Time-Sensitive Data**: Use `temporal` + `most_recent`
-3. **Multi-Source Integration**: Use `batch` + `merge_all` + `voting`
-4. **Medical/Research**: Use `credibility_weighted` for authoritative sources
-5. **Fraud/Security**: Use `graph_based` + `logical` + `expert_review`
-6. **Exact Matching Required**: Use `exact` strategy (IOCs, identifiers)
-

explorer/package.json

@@ -8,7 +8,8 @@
     "build": "tsc -b && vite build",
     "lint": "eslint .",
     "preview": "vite preview",
-    "test:graph-store": "node --test tests/graphStore.multi-edge.test.mjs"
+    "test:graph-store": "node --test tests/graphStore.multi-edge.test.mjs",
+    "test:graph-workspace": "node --import tsx --test tests/graphSceneState.display.test.ts"
   },
   "dependencies": {
     "@monaco-editor/react": "^4.7.0",
@@ -43,6 +44,7 @@
     "eslint-plugin-react-hooks": "^7.0.1",
     "eslint-plugin-react-refresh": "^0.5.2",
     "globals": "^17.4.0",
+    "tsx": "^4.21.0",
     "typescript": "~5.9.3",
     "typescript-eslint": "^8.57.0",
     "vite": "^5.4.0"

explorer/src/App.tsx

@@ -15,7 +15,7 @@ const EntityResolutionTab = lazy(() => import('./workspaces/EnrichWorkspace/Enti
 const KGOverviewTab = lazy(() => import('./workspaces/ManageWorkspace/KGOverviewTab').then((module) => ({ default: module.KGOverviewTab })));
 const OntologySummaryTab = lazy(() => import('./workspaces/ManageWorkspace/OntologySummaryTab').then((module) => ({ default: module.OntologySummaryTab })));
 
-type WorkspaceId = 'explore' | 'analyze' | 'decisions' | 'enrich' | 'manage';
+type WorkspaceId = 'welcome' | 'explore' | 'analyze' | 'decisions' | 'enrich' | 'manage';
 type ExploreView = 'graph' | 'vocabulary';
 type AnalyzeView = 'sparql' | 'reasoning';
 type EnrichView = 'import' | 'merge' | 'registry' | 'resolve';
@@ -311,14 +311,39 @@ function WorkspaceFallback() {
   return <div className="workspace-loading">Loading workspace…</div>;
 }
 
+function WelcomeScreen() {
+  return (
+    <div style={{
+      flex: 1,
+      display: 'flex',
+      flexDirection: 'column',
+      alignItems: 'center',
+      justifyContent: 'center',
+      gap: 12,
+      color: 'var(--text-muted)',
+    }}>
+      <h1 style={{ margin: 0, fontSize: 28, fontWeight: 700, color: 'var(--text-main)', letterSpacing: '-0.03em' }}>
+        Welcome to Semantica
+      </h1>
+      <p style={{ margin: 0, fontSize: 14 }}>
+        Select a workspace from the sidebar to get started.
+      </p>
+    </div>
+  );
+}
+
 export default function App() {
-  const [activeWorkspace, setActiveWorkspace] = useState<WorkspaceId>('explore');
+  const [activeWorkspace, setActiveWorkspace] = useState<WorkspaceId>('welcome');
   const [exploreView, setExploreView] = useState<ExploreView>('graph');
   const [analyzeView, setAnalyzeView] = useState<AnalyzeView>('reasoning');
   const [enrichView, setEnrichView] = useState<EnrichView>('import');
   const [manageView, setManageView] = useState<ManageView>('lineage');
 
   const renderWorkspace = () => {
+    if (activeWorkspace === 'welcome') {
+      return <WelcomeScreen />;
+    }
+
     if (activeWorkspace === 'explore') {
       return (
         <WorkspaceShell
@@ -451,7 +476,7 @@ export default function App() {
       <style>{shellStyles}</style>
       <div className="app-shell">
         <aside className="app-rail">
-          <div className="brand-pill" title="Semantica Knowledge Explorer">SKE</div>
+          <button className="brand-pill" title="Semantica Knowledge Explorer" onClick={() => setActiveWorkspace('welcome')} style={{ cursor: 'pointer', border: '1px solid rgba(127,208,255,0.18)' }}>SKE</button>
           {navItems.map(({ id, label, hint, icon: Icon }) => (
             <button
               key={id}

explorer/src/store/graphStore.ts

@@ -42,6 +42,10 @@ export interface NodeAttributes {
   haloColor?: string;
   labelVisibilityPolicy?: GraphLabelVisibilityPolicy;
   highlighted?: boolean;
+  communityId?: string;
+  isCommunityGroup?: boolean;
+  memberCount?: number;
+  anchorNodeId?: string | null;
 
   nodeType: string;
   content: string;
@@ -74,6 +78,12 @@ export interface EdgeAttributes {
   parallelIndex?: number;
   parallelCount?: number;
   familySize?: number;
+  rawEdgeIds?: string[];
+  isAggregated?: boolean;
+  aggregateCount?: number;
+  dominantEdgeType?: string;
+  representativeWeight?: number;
+  bundleKind?: "parallel" | "bidirectional" | "community";
   
  
   edgeType: string;

explorer/src/workspaces/GraphWorkspace/GraphInspectorPanel.tsx

@@ -2,6 +2,7 @@ import type { CSSProperties } from "react";
 import { Loader2 } from "lucide-react";
 import { graph } from "../../store/graphStore";
 import { GRAPH_THEME } from "./graphTheme";
+import type { GraphSelectedNodeKind } from "./types";
 
 export type LinkPrediction = {
   target: string;
@@ -14,10 +15,16 @@ export type PathResponse = {
   path: string[];
   edge_ids?: string[];
   total_weight: number;
+  hop_count: number;
+  distance_band: "direct" | "near" | "mid-range" | "distant";
 };
 
 export interface GraphInspectorPanelProps {
   nodeId: string;
+  inspectableNodeId?: string | null;
+  selectedNodeKind?: GraphSelectedNodeKind;
+  canActivateFocused?: boolean;
+  focusedUnavailableReason?: string | null;
   predictions: LinkPrediction[];
   predictionType: string;
   onPredictionTypeChange: (value: string) => void;
@@ -146,6 +153,10 @@ function PathFlowViz({
 
 export function GraphInspectorPanel({
   nodeId,
+  inspectableNodeId,
+  selectedNodeKind = "none",
+  canActivateFocused = false,
+  focusedUnavailableReason = null,
   predictions,
   predictionType,
   onPredictionTypeChange,
@@ -171,7 +182,38 @@ export function GraphInspectorPanel({
     );
   }
 
-  const attributes = graph.getNodeAttributes(nodeId) as {
+  const resolvedNodeId = inspectableNodeId && graph.hasNode(inspectableNodeId) ? inspectableNodeId : null;
+  const directlyInspectable = graph.hasNode(nodeId);
+  const effectiveNodeId = directlyInspectable ? nodeId : resolvedNodeId;
+  const actionNodeId = directlyInspectable ? nodeId : resolvedNodeId;
+  const groupedDisplaySelection = selectedNodeKind === "grouped" && !directlyInspectable;
+
+  if (!effectiveNodeId) {
+    return (
+      <aside style={{ padding: 24, display: "flex", flexDirection: "column", gap: 16 }}>
+        <div style={{ borderBottom: "1px solid rgba(88, 166, 255, 0.2)", paddingBottom: 16 }}>
+          <div style={{ display: "flex", alignItems: "center", gap: 10, marginBottom: 8 }}>
+            <span style={{ background: "#58a6ff", boxShadow: "0 0 10px rgba(88,166,255,0.45)", width: 8, height: 8, borderRadius: "50%" }} />
+            <span style={{ color: "#58a6ff", fontSize: 12, fontWeight: 700 }}>Selection</span>
+          </div>
+          <h3 style={{ margin: 0, color: "#fff", fontSize: 20, fontWeight: 700, wordBreak: "break-word" }}>
+            {nodeId}
+          </h3>
+          <div style={{ color: "#8b949e", fontSize: 12, marginTop: 6, fontFamily: "monospace", wordBreak: "break-all" }}>{nodeId}</div>
+        </div>
+        <div style={groupedSelectionNoticeStyle}>
+          <div style={{ color: "#dbe9f7", fontWeight: 600, marginBottom: 6 }}>Selected item is not directly inspectable in the current graph.</div>
+          <div style={{ color: "#8fa8c6", fontSize: 13, lineHeight: 1.6 }}>
+            {canActivateFocused
+              ? "Activate Focused mode to resolve this grouped selection to its canonical node."
+              : (focusedUnavailableReason ?? "Focused mode is unavailable for the current selection.")}
+          </div>
+        </div>
+      </aside>
+    );
+  }
+
+  const attributes = graph.getNodeAttributes(effectiveNodeId) as {
     color?: string;
     content?: string;
     label?: string;
@@ -194,12 +236,26 @@ export function GraphInspectorPanel({
       <div style={{ borderBottom: "1px solid rgba(88, 166, 255, 0.2)", paddingBottom: 16 }}>
         <div style={{ display: "flex", alignItems: "center", gap: 10, marginBottom: 8 }}>
           <span style={{ background: accentColor, boxShadow: `0 0 10px ${accentColor}`, width: 8, height: 8, borderRadius: "50%" }} />
-          <span style={{ color: accentColor, fontSize: 12, fontWeight: 700 }}>{attributes?.nodeType || "Entity"}</span>
+          <span style={{ color: accentColor, fontSize: 12, fontWeight: 700 }}>
+            {groupedDisplaySelection ? "Grouped Selection" : (attributes?.nodeType || "Entity")}
+          </span>
         </div>
         <h3 style={{ margin: 0, color: "#fff", fontSize: 20, fontWeight: 700, wordBreak: "break-word" }}>
-          {String(attributes?.label ?? nodeId)}
+          {String(attributes?.label ?? effectiveNodeId)}
         </h3>
-        <div style={{ color: "#8b949e", fontSize: 12, marginTop: 6, fontFamily: "monospace", wordBreak: "break-all" }}>{nodeId}</div>
+        <div style={{ color: "#8b949e", fontSize: 12, marginTop: 6, fontFamily: "monospace", wordBreak: "break-all" }}>
+          {groupedDisplaySelection ? nodeId : effectiveNodeId}
+        </div>
+        {groupedDisplaySelection ? (
+          <div style={groupedSelectionNoticeStyle}>
+            <div style={{ color: "#dbe9f7", fontWeight: 600, marginBottom: 6 }}>This grouped item stays display-level until you explicitly enter Focused mode.</div>
+            <div style={{ color: "#8fa8c6", fontSize: 13, lineHeight: 1.6 }}>
+              {canActivateFocused
+                ? `Canonical node available: ${effectiveNodeId}`
+                : (focusedUnavailableReason ?? "Focused mode is unavailable for the current selection.")}
+            </div>
+          </div>
+        ) : null}
         <div style={{ display: "flex", gap: 8, flexWrap: "wrap", marginTop: 12 }}>
           {attributes?.valid_from || attributes?.valid_until ? (
             <span style={subtleChipStyle}>temporal</span>
@@ -224,7 +280,7 @@ export function GraphInspectorPanel({
           <button
             style={{ ...actionButtonStyle, width: "100%", justifyContent: "center", opacity: isRunningPredictions ? 0.7 : 1 }}
             onClick={onRunPredictions}
-            disabled={isRunningPredictions}
+            disabled={isRunningPredictions || !actionNodeId}
           >
             {isRunningPredictions ? (
               <Loader2 size={14} className="animate-spin" style={{ marginRight: 6 }} />
@@ -232,10 +288,10 @@ export function GraphInspectorPanel({
             {isRunningPredictions ? "Running…" : "Run Link Prediction"}
           </button>
           <div style={{ display: "flex", gap: 8, flexWrap: "wrap" }}>
-            <button style={secondaryActionButtonStyle} onClick={() => onDownloadProvenance("json")}>
+            <button style={secondaryActionButtonStyle} onClick={() => onDownloadProvenance("json")} disabled={!actionNodeId}>
               Provenance JSON
             </button>
-            <button style={secondaryActionButtonStyle} onClick={() => onDownloadProvenance("markdown")}>
+            <button style={secondaryActionButtonStyle} onClick={() => onDownloadProvenance("markdown")} disabled={!actionNodeId}>
               Provenance MD
             </button>
           </div>
@@ -257,7 +313,7 @@ export function GraphInspectorPanel({
           placeholder="Target node ID"
           style={inputStyle}
         />
-        <button style={actionButtonStyle} onClick={onTracePath}>Trace Causal Path</button>
+        <button style={actionButtonStyle} onClick={onTracePath} disabled={!actionNodeId}>Trace Causal Path</button>
 
         {pathResult?.path?.length ? (
           <PathFlowViz
@@ -376,6 +432,14 @@ const inputStyle: CSSProperties = {
   boxShadow: "inset 0 1px 0 rgba(255,255,255,0.03)",
 };
 
+const groupedSelectionNoticeStyle: CSSProperties = {
+  marginTop: 12,
+  padding: "10px 12px",
+  background: "rgba(88,166,255,0.08)",
+  border: "1px solid rgba(88,166,255,0.2)",
+  borderRadius: 12,
+};
+
 const actionButtonStyle: CSSProperties = {
   background: "linear-gradient(135deg, rgba(24, 63, 133, 0.42), rgba(35, 85, 176, 0.28))",
   color: "#fff",

explorer/src/workspaces/GraphWorkspace/GraphRuntimeStage.tsx

@@ -3,6 +3,7 @@ import { forwardRef, useEffect, useImperativeHandle, useMemo, useRef, useState }
 import { batchMergeEdges, batchMergeNodes, clearGraph, graph, type EdgeAttributes, type NodeAttributes } from "../../store/graphStore";
 import { SigmaSceneAdapter } from "./SigmaSceneAdapter";
 import { createGraphLoadProgress } from "./graphLoading";
+import { resolveDisplayGraph } from "./graphSceneState";
 import {
   chooseColorAccessor,
   colorForNodeKey,
@@ -41,6 +42,7 @@ const STAGE_EFFECTS_STATE: GraphEffectsState = {
   lensMode: "neighborhood",
   effectQuality: "bounded",
 };
+const EMPTY_PATH: string[] = [];
 
 const socketProtocol = () => (window.location.protocol === "https:" ? "wss:" : "ws:");
 
@@ -67,6 +69,10 @@ function buildSelectedNodeState(nodeId: string): GraphSelectedNodeState | null {
     valid_until: attributes.valid_until ?? null,
     properties: attributes.properties ?? {},
     neighborCount: graph.neighbors(nodeId).length,
+    visibleNeighborCount: graph.neighbors(nodeId).length,
+    collapsedNeighborCount: 0,
+    isNeighborhoodCollapsed: false,
+    canCollapseNeighborhood: graph.neighbors(nodeId).length > 8,
   };
 }
 
@@ -113,6 +119,10 @@ export const GraphRuntimeStage = forwardRef<GraphStageHandle, GraphRuntimeStageP
     const prevActiveIdsRef = useRef<Set<string>>(new Set());
     const [graphVersion, setGraphVersion] = useState(0);
     const [runtimeLayoutSource, setRuntimeLayoutSource] = useState<GraphLayoutSource>(snapshot?.summary.layoutSource ?? "runtime");
+    const displayResult = useMemo(
+      () => resolveDisplayGraph(selectedNodeId, activePath, EMPTY_PATH, viewMode, { aggregationEnabled: true }),
+      [activePath, graphVersion, selectedNodeId, viewMode],
+    );
 
     const stageSignature = useMemo(() => (snapshot ? `${snapshot.fetchedAt}:${snapshot.summary.nodeCount}:${snapshot.summary.edgeCount}` : null), [snapshot]);
 
@@ -447,9 +457,16 @@ export const GraphRuntimeStage = forwardRef<GraphStageHandle, GraphRuntimeStageP
       <SigmaSceneAdapter
         ref={sceneRef}
         onNodeSelect={onNodeSelect}
+        graphVersion={graphVersion}
+        graphReady={Boolean(snapshot)}
+        displayGraph={displayResult.graph}
+        displayMeta={displayResult.meta}
+        displayState={displayResult.state}
         selectedEdgeId=""
         selectedNodeId={selectedNodeId}
+        focusedNodeId={viewMode === "focused" ? selectedNodeId : ""}
         activePath={activePath}
+        activePathEdgeIds={EMPTY_PATH}
         effectsState={STAGE_EFFECTS_STATE}
         isLayoutRunning={isLayoutRunning}
         onLayoutRunningChange={onLayoutRunningChange}

explorer/src/workspaces/GraphWorkspace/GraphWorkspace.tsx

@@ -1,5 +1,6 @@
 import { useCallback, useEffect, useMemo, useRef, useState } from "react";
 
+import type Graph from "graphology";
 import { batchMergeEdges, batchMergeNodes, graph } from "../../store/graphStore";
 import { logEvent } from "../../store/registryStore";
 import type { EdgeAttributes, NodeAttributes } from "../../store/graphStore";
@@ -11,6 +12,7 @@ import { useLoadGraph, useReloadGraph } from "./useLoadGraph";
 import { GraphLoadingOverlay } from "./GraphLoadingOverlay";
 import { createGraphLoadProgress, getGraphLoadTitle } from "./graphLoading";
 import { GRAPH_THEME, withAlpha } from "./graphTheme";
+import { checkGroupedViewAvailability, resolveDisplayGraph, resolveDisplayStateSnapshot, resolveGroupedDisplayNodeId, resolveGroupedDisplayStateSnapshot } from "./graphSceneState";
 import {
   type GraphPlugin,
   type GraphPluginActionRequest,
@@ -23,6 +25,7 @@ import type { LinkPrediction, PathResponse } from "./GraphInspectorPanel";
 import type { GraphSceneHandle, GraphSceneRuntime } from "./scene";
 import type {
   GraphAnalyticsSnapshot,
+  GraphDisplayStateSnapshot,
   GraphDiagnosticsSnapshot,
   GraphEffectToggle,
   GraphEffectsState,
@@ -30,6 +33,7 @@ import type {
   GraphLoadProgress,
   GraphLoadSummary,
   GraphSelectedEdgeState,
+  GraphSelectedNodeKind,
   GraphSelectedNodeState,
   GraphTemporalState,
   GraphViewMode,
@@ -89,7 +93,7 @@ const DEFAULT_EFFECTS_STATE: GraphEffectsState = {
   pathFlowEnabled: false,
   lensEnabled: false,
   temporalEmphasisEnabled: false,
-  semanticRegionsEnabled: true,
+  semanticRegionsEnabled: false,
   contoursEnabled: false,
   pathfindingEnabled: false,
   communitiesEnabled: false,
@@ -105,6 +109,15 @@ const LazyGraphInspectorPanel = lazy(() => import("./GraphInspectorPanel").then(
 const loadExplorationEffectsPlugin = () => import("./plugins/explorationEffectsPluginPhaseC").then((module) => module.explorationEffectsPluginPhaseC);
 const loadNeighborhoodPanelPlugin = () => import("./plugins/neighborhoodPanelPlugin").then((module) => module.neighborhoodPanelPlugin);
 const loadTemporalOverlayPlugin = () => import("./plugins/temporalOverlayPlugin").then((module) => module.temporalOverlayPlugin);
+const EMPTY_PATH: string[] = [];
+const DEBUG_GRAPH_WORKSPACE = import.meta.env.DEV;
+
+function debugGraphWorkspace(message: string, payload?: Record<string, unknown>) {
+  if (!DEBUG_GRAPH_WORKSPACE) {
+    return;
+  }
+  console.debug(`[GraphWorkspace] ${message}`, payload ?? {});
+}
 
 function useDebounce<T>(value: T, delay: number): T {
   const [debouncedValue, setDebouncedValue] = useState<T>(value);
@@ -494,7 +507,10 @@ function buildRealtimeEdgeAttributes(payload: {
   };
 }
 
-function buildSelectedNodeState(nodeId: string): GraphSelectedNodeState | null {
+function buildSelectedNodeState(
+  nodeId: string,
+  displayState: GraphDisplayStateSnapshot,
+): GraphSelectedNodeState | null {
   if (!nodeId || !graph.hasNode(nodeId)) {
     return null;
   }
@@ -519,31 +535,64 @@ function buildSelectedNodeState(nodeId: string): GraphSelectedNodeState | null {
     valid_until: attributes.valid_until ?? null,
     properties: attributes.properties ?? {},
     neighborCount: graph.neighbors(nodeId).length,
+    visibleNeighborCount: displayState.selectedVisibleNeighborIds.length,
+    collapsedNeighborCount: displayState.selectedCollapsedNeighborIds.length,
+    isNeighborhoodCollapsed: displayState.selectedCollapsedNeighborIds.length > 0,
+    canCollapseNeighborhood: graph.neighbors(nodeId).length > 8,
   };
 }
 
-function buildSelectedEdgeState(edgeId: string): GraphSelectedEdgeState | null {
-  if (!edgeId || !graph.hasEdge(edgeId)) {
+type FocusResolution = {
+  kind: GraphSelectedNodeKind;
+  resolvedNodeId: string | null;
+  reason: string | null;
+};
+
+function buildSelectedEdgeState(
+  edgeId: string,
+  displayGraph: typeof graph | Graph<NodeAttributes, EdgeAttributes>,
+): GraphSelectedEdgeState | null {
+  if (!edgeId || !displayGraph.hasEdge(edgeId)) {
     return null;
   }
 
-  const [sourceId, targetId] = graph.extremities(edgeId);
-  const attributes = graph.getEdgeAttributes(edgeId) as {
+  const [displaySourceId, displayTargetId] = displayGraph.extremities(edgeId);
+  const attributes = displayGraph.getEdgeAttributes(edgeId) as {
     edgeType?: string;
     weight?: number;
     properties?: Record<string, unknown>;
     familyId?: string;
+    rawEdgeIds?: string[];
+    isAggregated?: boolean;
+    aggregateCount?: number;
+    bundleKind?: "parallel" | "bidirectional" | "community";
+    dominantEdgeType?: string;
+    representativeWeight?: number;
   };
-  const sourceAttributes = graph.getNodeAttributes(sourceId) as { label?: string; content?: string };
-  const targetAttributes = graph.getNodeAttributes(targetId) as { label?: string; content?: string };
+  const rawEdgeIds = attributes.rawEdgeIds?.length ? attributes.rawEdgeIds.map((rawEdgeId) => String(rawEdgeId)) : [edgeId];
+  const primaryRawEdgeId = rawEdgeIds.find((rawEdgeId) => graph.hasEdge(rawEdgeId)) ?? rawEdgeIds[0];
+  let sourceId = displaySourceId;
+  let targetId = displayTargetId;
+  if (primaryRawEdgeId && graph.hasEdge(primaryRawEdgeId)) {
+    [sourceId, targetId] = graph.extremities(primaryRawEdgeId);
+  }
+  const sourceAttributes = graph.hasNode(sourceId)
+    ? (graph.getNodeAttributes(sourceId) as { label?: string; content?: string })
+    : ({ label: displaySourceId } as { label?: string; content?: string });
+  const targetAttributes = graph.hasNode(targetId)
+    ? (graph.getNodeAttributes(targetId) as { label?: string; content?: string })
+    : ({ label: displayTargetId } as { label?: string; content?: string });
   const properties = attributes.properties ?? {};
   const familyId = String(attributes.familyId || edgeId);
   let familySize = 0;
   let siblingCount = 0;
-  graph.forEachEdge((candidateEdgeId, candidateAttrs) => {
-    const edgeAttrs = candidateAttrs as { familyId?: string };
-    const [candidateSource, candidateTarget] = graph.extremities(candidateEdgeId);
-    if (String(edgeAttrs.familyId || candidateEdgeId) === familyId) {
+  rawEdgeIds.forEach((rawEdgeId) => {
+    if (!graph.hasEdge(rawEdgeId)) {
+      return;
+    }
+    const candidateAttrs = graph.getEdgeAttributes(rawEdgeId) as { familyId?: string };
+    const [candidateSource, candidateTarget] = graph.extremities(rawEdgeId);
+    if (String(candidateAttrs.familyId || rawEdgeId) === familyId) {
       familySize += 1;
     }
     if (candidateSource === sourceId && candidateTarget === targetId) {
@@ -551,6 +600,11 @@ function buildSelectedEdgeState(edgeId: string): GraphSelectedEdgeState | null {
     }
   });
 
+  if (attributes.isAggregated) {
+    siblingCount = rawEdgeIds.filter((rawEdgeId) => graph.hasEdge(rawEdgeId)).length;
+    familySize = Math.max(familySize, siblingCount);
+  }
+
   return {
     id: edgeId,
     familyId,
@@ -564,6 +618,12 @@ function buildSelectedEdgeState(edgeId: string): GraphSelectedEdgeState | null {
     provenanceCount: getProvenanceCount(properties),
     familySize,
     siblingCount,
+    isAggregated: Boolean(attributes.isAggregated),
+    aggregateCount: Number(attributes.aggregateCount ?? rawEdgeIds.length),
+    rawEdgeIds,
+    bundleKind: attributes.bundleKind ?? null,
+    dominantEdgeType: attributes.dominantEdgeType ?? null,
+    representativeWeight: Number(attributes.representativeWeight ?? attributes.weight ?? 1),
   };
 }
 
@@ -624,9 +684,15 @@ function collectPluginOverlays(
 
 export function GraphWorkspace() {
   const [selectedNodeId, setSelectedNodeId] = useState("");
+  const [focusedNodeId, setFocusedNodeId] = useState("");
+  const [lastGroupedSelectedNodeId, setLastGroupedSelectedNodeId] = useState("");
   const [selectedEdgeId, setSelectedEdgeId] = useState("");
   const [isLayoutRunning, setIsLayoutRunning] = useState(false);
-  const [viewMode, setViewMode] = useState<GraphViewMode>("focused");
+  const [graphReady, setGraphReady] = useState(false);
+  const [graphVersion, setGraphVersion] = useState(0);
+  const [viewMode, setViewMode] = useState<GraphViewMode>("full");
+  const [aggregationEnabled] = useState(true);
+  const [collapsedNeighborhoodNodeIds, setCollapsedNeighborhoodNodeIds] = useState<string[]>([]);
   const [searchQuery, setSearchQuery] = useState("");
   const [searchResults, setSearchResults] = useState<SearchResult[]>([]);
   const [searchError, setSearchError] = useState("");
@@ -663,15 +729,24 @@ export function GraphWorkspace() {
     focusedNodeId: "",
     activePath: [],
     activePathEdgeIds: [],
-    viewMode: "focused",
+    viewMode: "full",
     zoomTier: "overview",
     isLayoutRunning: false,
   });
   const reload = useReloadGraph();
 
+  const handleLoadProgress = useCallback((progress: GraphLoadProgress) => {
+    setLoadingProgress(progress);
+    if (progress.phase !== "ready" && progress.phase !== "stabilizing_layout") {
+      setGraphReady(false);
+    }
+  }, []);
+
   const { data: summary, isLoading, isFetching } = useLoadGraph({
     enabled: true,
     onGraphReady: (graphSummary) => {
+      setGraphReady(true);
+      setGraphVersion((current) => current + 1);
       setIsLayoutRunning(!graphSummary.layoutReady);
       if (settlingOverlayTimeoutRef.current !== null) {
         window.clearTimeout(settlingOverlayTimeoutRef.current);
@@ -700,7 +775,7 @@ export function GraphWorkspace() {
         settlingOverlayTimeoutRef.current = null;
       }, 900);
     },
-    onProgress: setLoadingProgress,
+    onProgress: handleLoadProgress,
   });
 
   useEffect(() => {
@@ -762,6 +837,7 @@ export function GraphWorkspace() {
           });
           prevActiveIdsRef.current = nextActiveIds;
           setActiveNodeCount(data.active_node_count);
+          setGraphVersion((current) => current + 1);
           sceneRef.current?.getRuntime()?.requestRender();
         });
       } catch (fetchError) {
@@ -777,16 +853,166 @@ export function GraphWorkspace() {
     };
   }, [debouncedTime, isLoading]);
 
+  const resolveNodeIdForFocusedMode = useCallback((
+    nodeId: string,
+    displayGraphCandidate?: GraphSceneRuntime["displayGraph"] | null,
+  ): FocusResolution => {
+    if (!nodeId) {
+      return {
+        kind: "none",
+        resolvedNodeId: null,
+        reason: "Select a node to inspect in Focused mode.",
+      };
+    }
+
+    if (graph.hasNode(nodeId)) {
+      return {
+        kind: "base",
+        resolvedNodeId: nodeId,
+        reason: null,
+      };
+    }
+
+    const currentDisplayGraph = displayGraphCandidate ?? pluginRuntimeRef.current?.displayGraph ?? graph;
+    if (currentDisplayGraph.hasNode(nodeId)) {
+      const displayAttrs = currentDisplayGraph.getNodeAttributes(nodeId) as NodeAttributes;
+      const communityGroup = displayAttrs.properties?.__communityGroup as
+        | {
+            anchorNodeId?: string | null;
+            sampleNodeIds?: string[];
+          }
+        | undefined;
+
+      const anchorNodeId = communityGroup?.anchorNodeId || communityGroup?.sampleNodeIds?.[0] || "";
+      if (anchorNodeId && graph.hasNode(anchorNodeId)) {
+        return {
+          kind: "grouped",
+          resolvedNodeId: anchorNodeId,
+          reason: null,
+        };
+      }
+
+      return {
+        kind: "grouped",
+        resolvedNodeId: null,
+        reason: "Focused mode is unavailable for this grouped selection.",
+      };
+    }
+
+    return {
+      kind: "unavailable",
+      resolvedNodeId: null,
+      reason: "Selected item is not available in the current graph.",
+    };
+  }, []);
+
+  const focusedSelectionResolution = useMemo(
+    () => resolveNodeIdForFocusedMode(selectedNodeId, pluginRuntimeRef.current?.displayGraph),
+    [pluginRuntimeVersion, resolveNodeIdForFocusedMode, selectedNodeId, viewMode],
+  );
+  const inspectableNodeId = focusedSelectionResolution.resolvedNodeId ?? "";
+  const canActivateFocusedMode = Boolean(focusedSelectionResolution.resolvedNodeId);
+  const { available: groupedViewAvailable, reason: groupedViewReason } = useMemo(
+    () => checkGroupedViewAvailability(),
+    [graphVersion],
+  );
+  const groupedDisplayCandidate = useMemo(
+    () => viewMode === "grouped"
+      ? resolveDisplayGraph("", EMPTY_PATH, EMPTY_PATH, "grouped", {
+          aggregationEnabled,
+          collapsedNeighborhoodNodeIds,
+        })
+      : null,
+    [viewMode, aggregationEnabled, collapsedNeighborhoodNodeIds, graphVersion],
+  );
+
+  const requestViewMode = useCallback((nextViewMode: GraphViewMode) => {
+    if (nextViewMode === "focused") {
+      const resolution = resolveNodeIdForFocusedMode(selectedNodeId, pluginRuntimeRef.current?.displayGraph);
+      if (!resolution.resolvedNodeId) {
+        return;
+      }
+
+      setFocusedNodeId(resolution.resolvedNodeId);
+      setSelectedNodeId(resolution.resolvedNodeId);
+      setViewMode("focused");
+      setIsLayoutRunning(false);
+      return;
+    }
+
+    if (nextViewMode === "grouped") {
+      if (!groupedViewAvailable) {
+        debugGraphWorkspace("grouped-view-unavailable", {
+          reason: groupedViewReason,
+          graphVersion,
+        });
+        return;
+      }
+
+      const groupedDisplayGraph = groupedDisplayCandidate?.graph
+        ?? resolveDisplayGraph("", EMPTY_PATH, EMPTY_PATH, "grouped", {
+          aggregationEnabled,
+          collapsedNeighborhoodNodeIds,
+        }).graph;
+      const nextGroupedSelection = [
+        lastGroupedSelectedNodeId,
+        selectedNodeId,
+        focusedNodeId,
+      ]
+        .map((candidateId) => resolveGroupedDisplayNodeId(groupedDisplayGraph, candidateId))
+        .find((candidateId): candidateId is string => Boolean(candidateId))
+        ?? "";
+
+      setFocusedNodeId("");
+      setSelectedNodeId(nextGroupedSelection);
+      if (nextGroupedSelection) {
+        setLastGroupedSelectedNodeId(nextGroupedSelection);
+      }
+      setViewMode("grouped");
+      setIsLayoutRunning(true);
+      return;
+    }
+
+    setFocusedNodeId("");
+    setSelectedNodeId((currentSelectedNodeId) => (
+      currentSelectedNodeId && graph.hasNode(currentSelectedNodeId) ? currentSelectedNodeId : ""
+    ));
+    setViewMode("full");
+  }, [
+    aggregationEnabled,
+    collapsedNeighborhoodNodeIds,
+    focusedNodeId,
+    graphVersion,
+    groupedDisplayCandidate,
+    groupedViewAvailable,
+    groupedViewReason,
+    lastGroupedSelectedNodeId,
+    resolveNodeIdForFocusedMode,
+    selectedNodeId,
+  ]);
+
   const focusNode = useCallback((nodeId: string) => {
-    setSelectedNodeId(nodeId);
+    if (!nodeId) {
+      return;
+    }
+
+    const currentDisplayGraph = pluginRuntimeRef.current?.displayGraph ?? graph;
+    const nextSelectedNodeId = nodeId;
+
+    if (!graph.hasNode(nodeId) && currentDisplayGraph.hasNode(nodeId)) {
+      setLastGroupedSelectedNodeId(nodeId);
+    }
+
+    setSelectedNodeId(nextSelectedNodeId);
     setSelectedEdgeId("");
     setPathResult(null);
     setSearchResults([]);
     setSearchError("");
-    if (nodeId) {
+    if (viewMode === "focused" && graph.hasNode(nextSelectedNodeId)) {
+      setFocusedNodeId(nextSelectedNodeId);
       setIsLayoutRunning(false);
     }
-  }, []);
+  }, [viewMode]);
 
   const handleEdgeSelect = useCallback((edgeId: string) => {
     setSelectedEdgeId(edgeId);
@@ -815,14 +1041,14 @@ export function GraphWorkspace() {
   }, [searchQuery]);
 
   const handleRunPredictions = useCallback(async () => {
-    if (!selectedNodeId) return;
+    if (!inspectableNodeId) return;
     setIsRunningPredictions(true);
     try {
       const response = await fetch("/api/enrich/links", {
         method: "POST",
         headers: { "Content-Type": "application/json" },
         body: JSON.stringify({
-          node_id: selectedNodeId,
+          node_id: inspectableNodeId,
           top_n: 6,
           candidate_type: predictionType || undefined,
           min_score: 0,
@@ -839,13 +1065,13 @@ export function GraphWorkspace() {
     } finally {
       setIsRunningPredictions(false);
     }
-  }, [predictionType, selectedNodeId]);
+  }, [inspectableNodeId, predictionType]);
 
   const handleTracePath = useCallback(async () => {
-    if (!selectedNodeId || !pathTargetId.trim()) return;
+    if (!inspectableNodeId || !pathTargetId.trim()) return;
     try {
       const response = await fetch(
-        `/api/graph/node/${encodeURIComponent(selectedNodeId)}/path?target=${encodeURIComponent(pathTargetId.trim())}&algorithm=dijkstra`
+        `/api/graph/node/${encodeURIComponent(inspectableNodeId)}/path?target=${encodeURIComponent(pathTargetId.trim())}&algorithm=dijkstra`
       );
       if (!response.ok) {
         throw new Error(`Path lookup failed with status ${response.status}`);
@@ -862,12 +1088,12 @@ export function GraphWorkspace() {
       console.error("[GraphWorkspace] path trace failed", pathError);
       setPathResult(null);
     }
-  }, [pathTargetId, selectedNodeId]);
+  }, [inspectableNodeId, pathTargetId]);
 
   const handleDownloadProvenance = useCallback(async (format: "json" | "markdown") => {
-    if (!selectedNodeId) return;
+    if (!inspectableNodeId) return;
     const suffix = format === "markdown" ? "markdown" : "json";
-    const response = await fetch(`/api/provenance/report?node_id=${encodeURIComponent(selectedNodeId)}&format=${suffix}`);
+    const response = await fetch(`/api/provenance/report?node_id=${encodeURIComponent(inspectableNodeId)}&format=${suffix}`);
     if (!response.ok) {
       throw new Error(`Provenance report failed with status ${response.status}`);
     }
@@ -875,12 +1101,12 @@ export function GraphWorkspace() {
     const url = window.URL.createObjectURL(blob);
     const anchor = document.createElement("a");
     anchor.href = url;
-    anchor.download = `${selectedNodeId}_provenance.${format === "markdown" ? "md" : "json"}`;
+    anchor.download = `${inspectableNodeId}_provenance.${format === "markdown" ? "md" : "json"}`;
     document.body.appendChild(anchor);
     anchor.click();
     window.URL.revokeObjectURL(url);
     document.body.removeChild(anchor);
-  }, [selectedNodeId]);
+  }, [inspectableNodeId]);
 
   useEffect(() => {
     const protocol = window.location.protocol === "https:" ? "wss:" : "ws:";
@@ -905,6 +1131,7 @@ export function GraphWorkspace() {
             },
           ]);
           logEvent("add-node", `Added node ${payload.label ?? payload.id}${payload.nodeType ? ` (${payload.nodeType})` : ""} via realtime ws`, { nodeId: payload.id, nodeType: payload.nodeType });
+          setGraphVersion((current) => current + 1);
           sceneRef.current?.getRuntime()?.requestRender();
         }
         if (eventType === "ADD_EDGE") {
@@ -918,6 +1145,7 @@ export function GraphWorkspace() {
             },
           ]);
           logEvent("add-edge", `Added edge ${payload.edgeType ?? payload.id} (${payload.source_id} → ${payload.target_id}) via realtime ws`, { edgeId: payload.id, edgeType: payload.edgeType, source: payload.source_id, target: payload.target_id });
+          setGraphVersion((current) => current + 1);
           sceneRef.current?.getRuntime()?.requestRender();
         }
       } catch (socketError) {
@@ -930,32 +1158,163 @@ export function GraphWorkspace() {
     };
   }, []);
 
+  useEffect(() => {
+    setCollapsedNeighborhoodNodeIds([]);
+    setFocusedNodeId("");
+    setLastGroupedSelectedNodeId("");
+  }, [summary?.edgeCount, summary?.nodeCount]);
+
+  const showLoadingOverlay = !graphReady && (isLoading || isFetching || Boolean(loadingProgress));
+  const showSettlingStatus = graphReady && loadingProgress?.phase === "stabilizing_layout";
+  const hasGraphContent = Boolean(summary?.nodeCount);
+  const activePath = pathResult?.path ?? EMPTY_PATH;
+  const activePathEdgeIds = pathResult?.edge_ids ?? EMPTY_PATH;
+  const structuralSelectedNodeId = useMemo(() => {
+    if (viewMode === "focused") {
+      return focusedNodeId && graph.hasNode(focusedNodeId) ? focusedNodeId : "";
+    }
+    if (!selectedNodeId || !graph.hasNode(selectedNodeId)) {
+      return "";
+    }
+    return collapsedNeighborhoodNodeIds.includes(selectedNodeId) ? selectedNodeId : "";
+  }, [collapsedNeighborhoodNodeIds, focusedNodeId, selectedNodeId, viewMode]);
+  const structuralActivePath = structuralSelectedNodeId ? activePath : EMPTY_PATH;
+  const structuralActivePathEdgeIds = structuralSelectedNodeId ? activePathEdgeIds : EMPTY_PATH;
+  const displayResult = useMemo(
+    () => (
+      viewMode === "grouped"
+        ? (groupedDisplayCandidate ?? resolveDisplayGraph("", EMPTY_PATH, EMPTY_PATH, "grouped", {
+            aggregationEnabled,
+            collapsedNeighborhoodNodeIds,
+          }))
+        : resolveDisplayGraph(structuralSelectedNodeId, structuralActivePath, structuralActivePathEdgeIds, viewMode, {
+            aggregationEnabled,
+            collapsedNeighborhoodNodeIds,
+          })
+    ),
+    [
+      aggregationEnabled,
+      collapsedNeighborhoodNodeIds,
+      groupedDisplayCandidate,
+      structuralActivePath,
+      structuralActivePathEdgeIds,
+      structuralSelectedNodeId,
+      viewMode,
+    ],
+  );
+  const displayState = useMemo(
+    () => (
+      viewMode === "grouped"
+        ? resolveGroupedDisplayStateSnapshot(displayResult.graph, selectedNodeId, {
+            groupedViewAvailable,
+            groupedViewReason,
+            selectedNodeKind: focusedSelectionResolution.kind,
+            resolvedFocusedNodeId: focusedSelectionResolution.resolvedNodeId,
+            focusedUnavailableReason: focusedSelectionResolution.reason,
+          })
+        : resolveDisplayStateSnapshot(selectedNodeId, activePath, viewMode, {
+            aggregationEnabled,
+            collapsedNeighborhoodNodeIds,
+            groupedViewAvailable,
+            groupedViewReason,
+            selectedNodeKind: focusedSelectionResolution.kind,
+            resolvedFocusedNodeId: focusedSelectionResolution.resolvedNodeId,
+            focusedUnavailableReason: focusedSelectionResolution.reason,
+          })
+    ),
+    [
+      activePath,
+      aggregationEnabled,
+      collapsedNeighborhoodNodeIds,
+      displayResult.graph,
+      focusedSelectionResolution.kind,
+      focusedSelectionResolution.reason,
+      focusedSelectionResolution.resolvedNodeId,
+      groupedViewAvailable,
+      groupedViewReason,
+      selectedNodeId,
+      viewMode,
+    ],
+  );
+  const displayMeta = displayResult.meta;
+  useEffect(() => {
+    if (viewMode === "grouped" && !groupedViewAvailable) {
+      debugGraphWorkspace("grouped-view-reset-to-full", {
+        reason: groupedViewReason,
+        graphVersion,
+      });
+      setViewMode("full");
+      setFocusedNodeId("");
+      setSelectedNodeId((currentSelectedNodeId) => (
+        currentSelectedNodeId && graph.hasNode(currentSelectedNodeId) ? currentSelectedNodeId : ""
+      ));
+    }
+  }, [graphVersion, groupedViewAvailable, groupedViewReason, viewMode]);
+  useEffect(() => {
+    if (viewMode === "focused" && (!focusedNodeId || !graph.hasNode(focusedNodeId))) {
+      setViewMode("full");
+      setFocusedNodeId("");
+    }
+  }, [focusedNodeId, graphVersion, viewMode]);
+  const previousDisplayGraphRef = useRef(displayResult.graph);
+  const previousDisplayStateRef = useRef(displayState);
+  useEffect(() => {
+    const graphRebuilt = previousDisplayGraphRef.current !== displayResult.graph;
+    const displayStateChanged = previousDisplayStateRef.current !== displayState;
+    debugGraphWorkspace("display-state-derived", {
+      selectedNodeId,
+      structuralSelectedNodeId,
+      viewMode,
+      graphRebuilt,
+      displayStateChanged,
+      aggregationEnabled,
+      collapsedNeighborhoodActive: Boolean(structuralSelectedNodeId && collapsedNeighborhoodNodeIds.includes(structuralSelectedNodeId)),
+    });
+    previousDisplayGraphRef.current = displayResult.graph;
+    previousDisplayStateRef.current = displayState;
+  }, [
+    aggregationEnabled,
+    collapsedNeighborhoodNodeIds,
+    displayResult.graph,
+    displayState,
+    selectedNodeId,
+    structuralSelectedNodeId,
+    viewMode,
+  ]);
   const focusedSummary = useMemo(() => {
     if (!selectedNodeId || !graph.hasNode(selectedNodeId)) {
+      if (viewMode === "grouped") {
+        return displayState.groupedViewAvailable
+          ? "Communities compressed into grouped structure view"
+          : (displayState.groupedViewReason ?? "Grouped view is unavailable for the current graph");
+      }
       return null;
     }
 
     const localNeighborCount = graph.neighbors(selectedNodeId).length;
     if (viewMode === "focused") {
-      const visibleNeighbors = Math.min(localNeighborCount, 16);
+      const visibleNeighbors = displayState.selectedVisibleNeighborIds.length || Math.min(localNeighborCount, 16);
       return `${visibleNeighbors + 1} nodes in focused view`;
     }
 
-    return `${localNeighborCount} direct neighbors highlighted`;
-  }, [selectedNodeId, viewMode]);
+    if (viewMode === "grouped") {
+      return "Grouped structure view with direct community drill-in";
+    }
 
-  const showLoadingOverlay = isLoading || isFetching || loadingProgress?.phase === "stabilizing_layout";
-  const hasGraphContent = Boolean(summary?.nodeCount);
-  const activePath = pathResult?.path ?? [];
-  const activePathEdgeIds = pathResult?.edge_ids ?? [];
+    if (displayState.selectedCollapsedNeighborIds.length > 0) {
+      return `${displayState.selectedVisibleNeighborIds.length} visible neighbors, ${displayState.selectedCollapsedNeighborIds.length} collapsed`;
+    }
+
+    return `${localNeighborCount} direct neighbors highlighted`;
+  }, [displayState, selectedNodeId, viewMode]);
   const graphSummary = summary as GraphLoadSummary | null;
   const selectedNodeState = useMemo(
-    () => buildSelectedNodeState(selectedNodeId),
-    [selectedNodeId, summary?.nodeCount, summary?.edgeCount],
+    () => buildSelectedNodeState(selectedNodeId, displayState),
+    [displayState, selectedNodeId, summary?.nodeCount, summary?.edgeCount],
   );
   const selectedEdgeState = useMemo(
-    () => buildSelectedEdgeState(selectedEdgeId),
-    [selectedEdgeId, summary?.nodeCount, summary?.edgeCount],
+    () => buildSelectedEdgeState(selectedEdgeId, displayResult.graph),
+    [displayResult.graph, selectedEdgeId, summary?.nodeCount, summary?.edgeCount],
   );
   const temporalState = useMemo(
     () => ({
@@ -1061,7 +1420,21 @@ export function GraphWorkspace() {
         focusNode(action.nodeId);
         return;
       case "setViewMode":
-        setViewMode(action.viewMode);
+        requestViewMode(action.viewMode);
+        return;
+      case "collapseNeighborhood":
+        if (!selectedNodeId) {
+          return;
+        }
+        setCollapsedNeighborhoodNodeIds((current) => (
+          current.includes(selectedNodeId) ? current : [...current, selectedNodeId]
+        ));
+        return;
+      case "expandNeighborhood":
+        if (!selectedNodeId) {
+          return;
+        }
+        setCollapsedNeighborhoodNodeIds((current) => current.filter((nodeId) => nodeId !== selectedNodeId));
         return;
       case "toggleEffect":
         setEffectToggle(action.effect, (current) => !current);
@@ -1099,7 +1472,7 @@ export function GraphWorkspace() {
         setActiveDockPanelId((previous) => (previous === action.panelId ? null : previous));
         return;
     }
-  }, [focusNode, setEffectToggle]);
+  }, [focusNode, requestViewMode, selectedNodeId, setEffectToggle]);
 
   const diagnosticsSnapshot = useMemo<GraphDiagnosticsSnapshot | null>(() => {
     if (!GRAPH_THEME.effects.diagnostics.enabledInDev || !graphDiagnosticsState) {
@@ -1139,9 +1512,11 @@ export function GraphWorkspace() {
     getEffectsState: () => effectsState,
     getDiagnosticsSnapshot: () => diagnosticsSnapshot,
     getAnalyticsSnapshot: () => graphAnalyticsState,
+    getDisplayState: () => displayState,
     isPanelOpen: (panelId: string) => Boolean(pluginPanelState[panelId]),
     dispatchAction: handlePluginAction,
   }), [
+    displayState,
     graphAnalyticsState,
     diagnosticsSnapshot,
     effectsState,
@@ -1267,23 +1642,58 @@ export function GraphWorkspace() {
   const coreToolbarGroups = useMemo<GraphToolbarGroup[]>(() => {
     const groups: GraphToolbarGroup[] = [];
 
-    if (selectedNodeId) {
+    if (hasGraphContent) {
       groups.push({
         id: "view-mode",
         items: [
-          {
-            id: "view-focused",
-            label: "Focused",
-            title: "Inspect the selected node in a focused local graph",
-            active: viewMode === "focused",
-            onClick: () => setViewMode("focused"),
-          },
           {
             id: "view-full",
             label: "Full Graph",
             title: "Return to the full graph context",
             active: viewMode === "full",
-            onClick: () => setViewMode("full"),
+            onClick: () => requestViewMode("full"),
+          },
+          {
+            id: "view-grouped",
+            label: "Grouped View",
+            title: displayState.groupedViewAvailable
+              ? "Compress dense structure into detected communities"
+              : (displayState.groupedViewReason ?? "Grouped view is unavailable until communities can be detected"),
+            active: viewMode === "grouped",
+            disabled: !displayState.groupedViewAvailable,
+            onClick: () => requestViewMode("grouped"),
+          },
+          {
+            id: "view-focused",
+            label: "Focused",
+            title: canActivateFocusedMode
+              ? "Inspect the selected node in a focused local graph"
+              : (focusedSelectionResolution.reason ?? "Focused mode is unavailable for the current selection"),
+            active: viewMode === "focused",
+            disabled: viewMode !== "focused" && !canActivateFocusedMode,
+            onClick: () => requestViewMode("focused"),
+          },
+        ],
+      });
+    }
+
+    if (selectedNodeState) {
+      groups.push({
+        id: "local-structure",
+        items: [
+          {
+            id: "collapse-neighborhood",
+            label: "Collapse Neighborhood",
+            title: "Hide lower-priority fanout around the selected node",
+            disabled: !selectedNodeState.canCollapseNeighborhood || selectedNodeState.isNeighborhoodCollapsed,
+            onClick: () => handlePluginAction({ type: "collapseNeighborhood" }),
+          },
+          {
+            id: "expand-neighborhood",
+            label: "Expand Neighborhood",
+            title: "Restore the collapsed local neighborhood",
+            disabled: !selectedNodeState.isNeighborhoodCollapsed,
+            onClick: () => handlePluginAction({ type: "expandNeighborhood" }),
           },
         ],
       });
@@ -1304,25 +1714,13 @@ export function GraphWorkspace() {
           id: "zoom-in",
           label: "＋ Zoom In",
           title: "Zoom in (or scroll up on the canvas)",
-          onClick: () => {
-            const runtime = sceneRef.current?.getRuntime();
-            if (runtime?.renderer === "sigma") {
-              const camera = (runtime.scene as import("sigma").default).getCamera();
-              camera.animatedZoom({ duration: 200 });
-            }
-          },
+          onClick: () => sceneRef.current?.zoomIn(),
         },
         {
           id: "zoom-out",
           label: "－ Zoom Out",
           title: "Zoom out (or scroll down on the canvas)",
-          onClick: () => {
-            const runtime = sceneRef.current?.getRuntime();
-            if (runtime?.renderer === "sigma") {
-              const camera = (runtime.scene as import("sigma").default).getCamera();
-              camera.animatedUnzoom({ duration: 200 });
-            }
-          },
+          onClick: () => sceneRef.current?.zoomOut(),
         },
         {
           id: "fit-view",
@@ -1362,18 +1760,42 @@ export function GraphWorkspace() {
     }
 
     return groups;
-  }, [isLayoutRunning, pluginToolbarItems, reload, searchQuery, selectedNodeId, showLoadingOverlay, viewMode]);
+  }, [
+    canActivateFocusedMode,
+    displayState.groupedViewAvailable,
+    displayState.groupedViewReason,
+    handlePluginAction,
+    hasGraphContent,
+    isLayoutRunning,
+    pluginToolbarItems,
+    reload,
+    requestViewMode,
+    searchQuery,
+    canActivateFocusedMode,
+    focusedSelectionResolution.reason,
+    selectedNodeId,
+    selectedNodeState,
+    showLoadingOverlay,
+    viewMode,
+  ]);
 
   const sceneAdapterProps = {
     onNodeSelect: focusNode,
     onEdgeSelect: handleEdgeSelect,
+    graphVersion,
+    graphReady,
+    displayGraph: displayResult.graph,
+    displayMeta,
+    displayState,
     selectedNodeId,
+    focusedNodeId,
     selectedEdgeId,
     activePath,
     activePathEdgeIds,
     effectsState,
     temporalState,
     isLayoutRunning,
+    layoutSource: graphSummary?.layoutSource,
     viewMode,
     showFitViewButton: false,
     pluginOverlays: pluginOverlays.map((overlay) => overlay.element),
@@ -1395,7 +1817,7 @@ export function GraphWorkspace() {
             <div style={{ display: "flex", flexDirection: "column", gap: 12 }}>
               <div className="explore-toolbar">
                 <div style={{ display: "flex", gap: 8, flexWrap: "wrap", alignItems: "center" }}>
-                  {showLoadingOverlay && loadingProgress ? (
+                  {(showLoadingOverlay || showSettlingStatus) && loadingProgress ? (
                     <MetricChip>{getGraphLoadTitle(loadingProgress.phase)}</MetricChip>
                   ) : null}
                   {summary ? (
@@ -1504,8 +1926,20 @@ export function GraphWorkspace() {
 
                   <div style={{ display: "flex", gap: 8, flexWrap: "wrap" }}>
                     <MetricChip tone="warm">weight {selectedEdgeState.weight.toFixed(2)}</MetricChip>
-                    <MetricChip>{selectedEdgeState.siblingCount} parallel lane{selectedEdgeState.siblingCount === 1 ? "" : "s"}</MetricChip>
+                    {selectedEdgeState.isAggregated ? (
+                      <MetricChip tone="success">
+                        {selectedEdgeState.aggregateCount} bundled edge{selectedEdgeState.aggregateCount === 1 ? "" : "s"}
+                      </MetricChip>
+                    ) : (
+                      <MetricChip>{selectedEdgeState.siblingCount} parallel lane{selectedEdgeState.siblingCount === 1 ? "" : "s"}</MetricChip>
+                    )}
                     <MetricChip>{selectedEdgeState.familySize} family member{selectedEdgeState.familySize === 1 ? "" : "s"}</MetricChip>
+                    {selectedEdgeState.bundleKind ? (
+                      <MetricChip>{selectedEdgeState.bundleKind} bundle</MetricChip>
+                    ) : null}
+                    {selectedEdgeState.dominantEdgeType ? (
+                      <MetricChip>{selectedEdgeState.dominantEdgeType}</MetricChip>
+                    ) : null}
                     {selectedEdgeState.provenanceCount > 0 ? (
                       <MetricChip>{selectedEdgeState.provenanceCount} provenance fields</MetricChip>
                     ) : null}
@@ -1608,6 +2042,10 @@ export function GraphWorkspace() {
                   <Suspense fallback={<div style={inspectorFallbackStyle}>Loading inspector…</div>}>
                     <LazyGraphInspectorPanel
                       nodeId={selectedNodeId}
+                      inspectableNodeId={inspectableNodeId || null}
+                      selectedNodeKind={displayState.selectedNodeKind}
+                      canActivateFocused={canActivateFocusedMode}
+                      focusedUnavailableReason={displayState.focusedUnavailableReason}
                       predictions={predictions}
                       predictionType={predictionType}
                       onPredictionTypeChange={setPredictionType}

explorer/src/workspaces/GraphWorkspace/GraphWorkspaceShell.tsx

@@ -33,6 +33,8 @@ type LinkPrediction = {
 type PathResponse = {
   path: GraphPath;
   total_weight: number;
+  hop_count: number;
+  distance_band: "direct" | "near" | "mid-range" | "distant";
 };
 
 type TemporalBounds = {
@@ -137,6 +139,10 @@ function toSelectedNodeState(node: ApiNode, neighborCount: number, fallbackColor
     valid_until: node.valid_until ?? null,
     properties: node.properties ?? {},
     neighborCount,
+    visibleNeighborCount: neighborCount,
+    collapsedNeighborCount: 0,
+    isNeighborhoodCollapsed: false,
+    canCollapseNeighborhood: neighborCount > 8,
   };
 }
 
@@ -425,6 +431,10 @@ export function GraphWorkspaceShell() {
           valid_until: null,
           properties: searchNode.properties ?? {},
           neighborCount: 0,
+          visibleNeighborCount: 0,
+          collapsedNeighborCount: 0,
+          isNeighborhoodCollapsed: false,
+          canCollapseNeighborhood: false,
         }
       : null;
   }, [neighborCountMap, searchResults, selectedNodeId, selectedNodeState, snapshot]);
@@ -553,6 +563,19 @@ export function GraphWorkspaceShell() {
     return `${visibleSelectedNode.neighborCount} direct neighbors highlighted`;
   }, [viewMode, visibleSelectedNode]);
 
+  const requestViewMode = useCallback((nextViewMode: GraphViewMode) => {
+    if (nextViewMode === "focused") {
+      if (!selectedNodeId) {
+        return;
+      }
+      setViewMode("focused");
+      setIsLayoutRunning(false);
+      return;
+    }
+
+    setViewMode("full");
+  }, [selectedNodeId]);
+
   const showLoadingOverlay =
     isLoading
     || isFetching
@@ -629,8 +652,8 @@ export function GraphWorkspaceShell() {
               <div className="graph-toggle-cluster">
               {selectedNodeId ? (
                   <>
-                    <button onClick={() => setViewMode("focused")} style={{ ...actionButtonStyle, background: viewMode === "focused" ? "rgba(31, 111, 235, 0.38)" : actionButtonStyle.background, borderColor: viewMode === "focused" ? "rgba(127, 208, 255, 0.42)" : "rgba(88, 166, 255, 0.2)" }}>Focused View</button>
-                    <button onClick={() => setViewMode("full")} style={{ ...actionButtonStyle, background: viewMode === "full" ? "rgba(31, 111, 235, 0.38)" : actionButtonStyle.background, borderColor: viewMode === "full" ? "rgba(127, 208, 255, 0.42)" : "rgba(88, 166, 255, 0.2)" }}>Full Graph</button>
+                    <button onClick={() => requestViewMode("focused")} style={{ ...actionButtonStyle, background: viewMode === "focused" ? "rgba(31, 111, 235, 0.38)" : actionButtonStyle.background, borderColor: viewMode === "focused" ? "rgba(127, 208, 255, 0.42)" : "rgba(88, 166, 255, 0.2)" }}>Focused View</button>
+                    <button onClick={() => requestViewMode("full")} style={{ ...actionButtonStyle, background: viewMode === "full" ? "rgba(31, 111, 235, 0.38)" : actionButtonStyle.background, borderColor: viewMode === "full" ? "rgba(127, 208, 255, 0.42)" : "rgba(88, 166, 255, 0.2)" }}>Full Graph</button>
                   </>
                 ) : (
                   <span style={{ color: "#7f95b3", fontSize: 12 }}>Select a node to switch graph views</span>

explorer/src/workspaces/GraphWorkspace/SigmaSceneAdapter.tsx

@@ -24,6 +24,8 @@ export const SigmaSceneAdapter = forwardRef<GraphSceneHandle, GraphSceneProps>(
     useImperativeHandle(ref, () => ({
       fitView: () => canvasRef.current?.fitView(),
       focusNode: (nodeId: string) => canvasRef.current?.focusNode(nodeId),
+      zoomIn: () => canvasRef.current?.zoomIn(),
+      zoomOut: () => canvasRef.current?.zoomOut(),
       getRuntime: () => runtimeRef.current,
       setLayoutRunning: onLayoutRunningChange
         ? (running: boolean) => {

explorer/src/workspaces/GraphWorkspace/behaviors/focusCameraBehavior.ts

@@ -5,11 +5,21 @@ export const focusCameraBehavior: GraphBehavior = {
   attach: () => {},
   detach: () => {},
   performAction: (context, action) => {
-    if (action.type !== "focusNode") {
-      return false;
+    if (action.type === "focusNode") {
+      context.focusNodeInView(action.nodeId);
+      return true;
     }
 
-    context.focusNodeInView(action.nodeId);
-    return true;
+    if (action.type === "centerSelection") {
+      context.centerSelectionInView(action.nodeId);
+      return true;
+    }
+
+    if (action.type === "centerGroupedSelection") {
+      context.centerGroupedSelectionInView(action.nodeId);
+      return true;
+    }
+
+    return false;
   },
 };

explorer/src/workspaces/GraphWorkspace/behaviors/searchFocusBehavior.ts

@@ -1,23 +1,36 @@
 import type { GraphBehavior } from "./types";
 
 export function createSearchFocusBehavior(): GraphBehavior {
-  let lastFocusedNodeId = "";
+  let lastSelectedNodeId = "";
+  let lastViewMode = "";
 
   return {
     id: "search-focus",
     attach: () => {},
     detach: () => {
-      lastFocusedNodeId = "";
+      lastSelectedNodeId = "";
+      lastViewMode = "";
     },
     onStateChange: (context, interactionState) => {
-      const nextFocusedNodeId = interactionState.focusedNodeId;
-      if (!nextFocusedNodeId || nextFocusedNodeId === lastFocusedNodeId) {
-        lastFocusedNodeId = nextFocusedNodeId;
+      const nextSelectedNodeId = interactionState.selectedNodeId;
+      const nextViewMode = interactionState.viewMode;
+      if (nextViewMode !== lastViewMode) {
+        lastViewMode = nextViewMode;
+        lastSelectedNodeId = nextSelectedNodeId;
+        return;
+      }
+      if (!nextSelectedNodeId || nextSelectedNodeId === lastSelectedNodeId) {
+        lastSelectedNodeId = nextSelectedNodeId;
+        lastViewMode = nextViewMode;
         return;
       }
 
-      lastFocusedNodeId = nextFocusedNodeId;
-      context.dispatchAction({ type: "focusNode", nodeId: nextFocusedNodeId });
+      lastSelectedNodeId = nextSelectedNodeId;
+      lastViewMode = nextViewMode;
+      context.dispatchAction({
+        type: nextViewMode === "grouped" ? "centerGroupedSelection" : "centerSelection",
+        nodeId: nextSelectedNodeId,
+      });
     },
   };
 }

explorer/src/workspaces/GraphWorkspace/behaviors/types.ts

@@ -6,7 +6,9 @@ import type { GraphCameraState, GraphInteractionState } from "../types";
 
 export type GraphBehaviorActionRequest =
   | { type: "fitView" }
-  | { type: "focusNode"; nodeId: string };
+  | { type: "focusNode"; nodeId: string }
+  | { type: "centerSelection"; nodeId: string }
+  | { type: "centerGroupedSelection"; nodeId: string };
 
 export interface GraphBehaviorContext {
   sigma: Sigma;
@@ -17,6 +19,8 @@ export interface GraphBehaviorContext {
   onNodeSelectionChange: (nodeId: string) => void;
   onEdgeSelectionChange: (edgeId: string) => void;
   focusNodeInView: (nodeId: string) => void;
+  centerSelectionInView: (nodeId: string) => void;
+  centerGroupedSelectionInView: (nodeId: string) => void;
   fitCurrentView: () => void;
   dispatchAction: (action: GraphBehaviorActionRequest) => void;
 }

explorer/src/workspaces/GraphWorkspace/behaviors/viewModeSwitchBehavior.ts

@@ -1,7 +1,8 @@
 import type { GraphBehavior } from "./types";
+import type { GraphViewMode } from "../types";
 
 export function createViewModeSwitchBehavior(): GraphBehavior {
-  let lastViewMode: "focused" | "full" | null = null;
+  let lastViewMode: GraphViewMode | null = null;
 
   return {
     id: "view-mode-switch",
@@ -15,9 +16,10 @@ export function createViewModeSwitchBehavior(): GraphBehavior {
       }
 
       lastViewMode = interactionState.viewMode;
+      const nextFocusedNodeId = interactionState.focusedNodeId;
 
-      if (interactionState.focusedNodeId) {
-        context.dispatchAction({ type: "focusNode", nodeId: interactionState.focusedNodeId });
+      if (interactionState.viewMode === "focused" && nextFocusedNodeId) {
+        context.dispatchAction({ type: "focusNode", nodeId: nextFocusedNodeId });
         return;
       }
 

explorer/src/workspaces/GraphWorkspace/graphTheme.ts

@@ -9,6 +9,7 @@ export type GraphBadgeKind = "inferred" | "temporal" | "provenance";
 
 type GraphNodeColorMode = "base" | "selected" | "hovered" | "path" | "muted";
 type GraphEdgeColorMode = "overview" | "backbone" | "structure" | "inspection" | "hover" | "path" | "focus" | "muted";
+const IS_DEV = Boolean((import.meta as { env?: { DEV?: boolean } }).env?.DEV);
 
 export interface GraphTheme {
   palette: {
@@ -190,6 +191,35 @@ export interface GraphTheme {
   motion: {
     cameraMs: number;
   };
+  grouped: {
+    initialLayout: {
+      innerRadius: number;
+      ringSpacing: number;
+      minNodeSpacing: number;
+      nodePadding: number;
+      overlapIterations: number;
+      primaryLabelCount: number;
+    };
+    style: {
+      nodeSizeScale: number;
+      nodeBorderBoost: number;
+      fillAlpha: number;
+      shellAlpha: number;
+      edgeSizeScale: number;
+      edgeAlpha: number;
+      glowAlpha: number;
+      edgeVisibilityRatio: number;
+      topIncidentEdges: number;
+    };
+    layout: {
+      iterations: number;
+      gravity: number;
+      scalingRatio: number;
+      edgeWeightInfluence: number;
+      slowDown: number;
+      settleMs: number;
+    };
+  };
   effects: {
     pathPulse: {
       minZoomTier: GraphZoomTier;
@@ -305,10 +335,10 @@ export const GRAPH_THEME: GraphTheme = {
   zoomTiers: {
     overview: {
       maxRatio: Number.POSITIVE_INFINITY,
-      nodeScale: 0.88,
-      labelThreshold: 0.92,
-      labelBudget: 28,
-      edgePriorityThreshold: 0.55,
+      nodeScale: 0.72,
+      labelThreshold: 0.995,
+      labelBudget: 4,
+      edgePriorityThreshold: 0.72,
       arrowPriorityThreshold: Number.POSITIVE_INFINITY,
       edgeSizeScale: 0.62,
       showBadges: false,
@@ -317,21 +347,21 @@ export const GRAPH_THEME: GraphTheme = {
     },
     structure: {
       maxRatio: 1.2,
-      nodeScale: 1.02,
-      labelThreshold: 0.82,
-      labelBudget: 60,
-      edgePriorityThreshold: 0.3,
-      arrowPriorityThreshold: 0.65,
-      edgeSizeScale: 1.05,
-      showBadges: true,
-      showCurves: true,
-      showContextualArrows: true,
+      nodeScale: 0.94,
+      labelThreshold: 0.93,
+      labelBudget: 18,
+      edgePriorityThreshold: 0.4,
+      arrowPriorityThreshold: 0.75,
+      edgeSizeScale: 0.92,
+      showBadges: false,
+      showCurves: false,
+      showContextualArrows: false,
     },
     inspection: {
       maxRatio: 0.5,
-      nodeScale: 1.08,
-      labelThreshold: 0.6,
-      labelBudget: 120,
+      nodeScale: 1,
+      labelThreshold: 0.8,
+      labelBudget: 40,
       edgePriorityThreshold: 0,
       arrowPriorityThreshold: 0.45,
       edgeSizeScale: 1.18,
@@ -341,7 +371,7 @@ export const GRAPH_THEME: GraphTheme = {
     },
   },
   labels: {
-    forceVisibleStates: ["hovered", "selected", "neighbor", "path"],
+    forceVisibleStates: ["hovered", "selected", "path"],
     policies: {
       none: { minZoomTier: "inspection" },
       priority: { minZoomTier: "overview" },
@@ -391,26 +421,26 @@ export const GRAPH_THEME: GraphTheme = {
   },
   nodes: {
     backgroundScale: 0.52,
-    mutedAlpha: 0.08,
+    mutedAlpha: 0.16,
     strokeHierarchy: {
       overview: { base: 0.05, emphasis: 0.34, muted: 0.02 },
       structure: { base: 1.05, emphasis: 1.45, muted: 0.55 },
       inspection: { base: 1.2, emphasis: 1.7, muted: 0.6 },
     },
     states: {
-      default: { color: "base", sizeMultiplier: 0.92, minSize: 3.5, forceLabel: false, zIndex: 0, borderBoost: -0.18 },
-      hovered: { color: "hovered", sizeMultiplier: 1.28, minSize: 13.5, forceLabel: true, zIndex: 4, borderBoost: 0.28 },
-      selected: { color: "selected", sizeMultiplier: 1.14, minSize: 11.5, forceLabel: true, zIndex: 3, borderBoost: 0.24 },
-      neighbor: { color: "base", sizeMultiplier: 0.96, minSize: 5.5, forceLabel: true, zIndex: 2, borderBoost: 0.04 },
-      path: { color: "path", sizeMultiplier: 1.08, minSize: 7.0, forceLabel: true, zIndex: 2, borderBoost: 0.12 },
-      inactive: { color: "muted", sizeMultiplier: 0.48, minSize: 1.8, forceLabel: false, zIndex: 0, borderBoost: -0.28 },
-      muted: { color: "muted", sizeMultiplier: 0.48, minSize: 1.8, forceLabel: false, zIndex: 0, borderBoost: -0.28 },
+      default: { color: "base", sizeMultiplier: 0.72, minSize: 0.68, forceLabel: false, zIndex: 0, borderBoost: -0.42 },
+      hovered: { color: "hovered", sizeMultiplier: 1.1, minSize: 10.8, forceLabel: true, zIndex: 4, borderBoost: 0.18 },
+      selected: { color: "selected", sizeMultiplier: 1.02, minSize: 9.4, forceLabel: true, zIndex: 3, borderBoost: 0.18 },
+      neighbor: { color: "base", sizeMultiplier: 0.78, minSize: 4.2, forceLabel: false, zIndex: 2, borderBoost: -0.12 },
+      path: { color: "path", sizeMultiplier: 0.97, minSize: 5.8, forceLabel: true, zIndex: 2, borderBoost: 0.06 },
+      inactive: { color: "muted", sizeMultiplier: 0.52, minSize: 0.58, forceLabel: false, zIndex: 0, borderBoost: -0.42 },
+      muted: { color: "muted", sizeMultiplier: 0.52, minSize: 0.58, forceLabel: false, zIndex: 0, borderBoost: -0.42 },
     },
     variants: {
       default: { sizeMultiplier: 1, borderBoost: 0, haloBoost: 0, badgeVisibleFrom: "inspection" },
-      temporal: { sizeMultiplier: 1.02, borderBoost: 0.12, haloBoost: 0.1, badgeKind: "temporal", badgeVisibleFrom: "structure" },
-      inferred: { sizeMultiplier: 1.05, borderBoost: 0.16, haloBoost: 0.14, badgeKind: "inferred", badgeVisibleFrom: "structure" },
-      provenance: { sizeMultiplier: 1.03, borderBoost: 0.14, haloBoost: 0.12, badgeKind: "provenance", badgeVisibleFrom: "structure" },
+      temporal: { sizeMultiplier: 1.02, borderBoost: 0.12, haloBoost: 0.1, badgeKind: "temporal", badgeVisibleFrom: "inspection" },
+      inferred: { sizeMultiplier: 1.05, borderBoost: 0.16, haloBoost: 0.14, badgeKind: "inferred", badgeVisibleFrom: "inspection" },
+      provenance: { sizeMultiplier: 1.03, borderBoost: 0.14, haloBoost: 0.12, badgeKind: "provenance", badgeVisibleFrom: "inspection" },
       selected: { sizeMultiplier: 1.06, borderBoost: 0.22, haloBoost: 0.16, badgeVisibleFrom: "overview" },
     },
     selectedRing: {
@@ -473,6 +503,35 @@ export const GRAPH_THEME: GraphTheme = {
   motion: {
     cameraMs: 380,
   },
+  grouped: {
+    initialLayout: {
+      innerRadius: 92,
+      ringSpacing: 138,
+      minNodeSpacing: 112,
+      nodePadding: 28,
+      overlapIterations: 18,
+      primaryLabelCount: 6,
+    },
+    style: {
+      nodeSizeScale: 0.9,
+      nodeBorderBoost: 0.42,
+      fillAlpha: 0.68,
+      shellAlpha: 0.28,
+      edgeSizeScale: 0.62,
+      edgeAlpha: 0.32,
+      glowAlpha: 0.14,
+      edgeVisibilityRatio: 0.18,
+      topIncidentEdges: 2,
+    },
+    layout: {
+      iterations: 18,
+      gravity: 0.06,
+      scalingRatio: 18,
+      edgeWeightInfluence: 0.08,
+      slowDown: 34,
+      settleMs: 1500,
+    },
+  },
   effects: {
     pathPulse: {
       minZoomTier: "structure",
@@ -530,7 +589,7 @@ export const GRAPH_THEME: GraphTheme = {
       maxGroups: 8,
     },
     diagnostics: {
-      enabledInDev: import.meta.env.DEV,
+      enabledInDev: IS_DEV,
     },
   },
 };

explorer/src/workspaces/GraphWorkspace/plugins/neighborhoodPanelPlugin.tsx

@@ -42,6 +42,7 @@ export const neighborhoodPanelPlugin: GraphPlugin = {
     }
 
     const selected = context.getSelectedNodeState();
+    const displayState = context.getDisplayState();
     if (!selected) {
       return {
         id: NEIGHBORHOOD_PANEL_ID,
@@ -82,6 +83,11 @@ export const neighborhoodPanelPlugin: GraphPlugin = {
         return left.label.localeCompare(right.label);
       })
       .slice(0, MAX_NEIGHBORS);
+    const hiddenNeighborCount = displayState.selectedCollapsedNeighborIds.length;
+    const aggregatedEdgeCount = context.displayGraph
+      .edges()
+      .map((edgeId) => context.displayGraph.getEdgeAttributes(edgeId) as { isAggregated?: boolean })
+      .filter((attrs) => attrs.isAggregated).length;
 
     return {
       id: NEIGHBORHOOD_PANEL_ID,
@@ -97,6 +103,34 @@ export const neighborhoodPanelPlugin: GraphPlugin = {
           <div style={summaryStyle}>
             {selected.neighborCount.toLocaleString()} direct neighbors in the full graph
           </div>
+          <div style={{ display: "flex", gap: 8, flexWrap: "wrap" }}>
+            <button
+              type="button"
+              onClick={() => context.dispatchAction({ type: "collapseNeighborhood" })}
+              disabled={!selected.canCollapseNeighborhood || selected.isNeighborhoodCollapsed}
+              style={controlButtonStyle}
+            >
+              Collapse Neighborhood
+            </button>
+            <button
+              type="button"
+              onClick={() => context.dispatchAction({ type: "expandNeighborhood" })}
+              disabled={!selected.isNeighborhoodCollapsed}
+              style={controlButtonStyle}
+            >
+              Expand Neighborhood
+            </button>
+          </div>
+          {hiddenNeighborCount > 0 ? (
+            <div style={summaryStyle}>
+              {hiddenNeighborCount.toLocaleString()} lower-priority neighbors are collapsed in the current view.
+            </div>
+          ) : null}
+          {aggregatedEdgeCount > 0 ? (
+            <div style={summaryStyle}>
+              {aggregatedEdgeCount.toLocaleString()} aggregated structural bundle{aggregatedEdgeCount === 1 ? "" : "s"} visible.
+            </div>
+          ) : null}
           {neighbors.length ? (
             <div style={{ display: "flex", flexDirection: "column", gap: 8 }}>
               {neighbors.map((neighbor) => (
@@ -159,6 +193,16 @@ const neighborButtonStyle: CSSProperties = {
   cursor: "pointer",
 };
 
+const controlButtonStyle: CSSProperties = {
+  padding: "7px 10px",
+  background: "rgba(255,255,255,0.03)",
+  border: "1px solid rgba(255,255,255,0.08)",
+  borderRadius: 10,
+  color: "#dce7f4",
+  cursor: "pointer",
+  fontSize: 12,
+};
+
 const swatchStyle: CSSProperties = {
   width: 10,
   height: 10,

explorer/src/workspaces/GraphWorkspace/plugins/types.ts

@@ -6,6 +6,7 @@ import type { GraphTheme } from "../graphTheme";
 import type { GraphSceneRuntime } from "../scene";
 import type {
   GraphAnalyticsSnapshot,
+  GraphDisplayStateSnapshot,
   GraphDiagnosticsSnapshot,
   GraphEffectsState,
   GraphEffectToggle,
@@ -31,6 +32,8 @@ export type GraphPluginActionRequest =
   | { type: "focusNode"; nodeId: string }
   | { type: "selectNode"; nodeId: string }
   | { type: "setViewMode"; viewMode: GraphViewMode }
+  | { type: "collapseNeighborhood" }
+  | { type: "expandNeighborhood" }
   | { type: "toggleEffect"; effect: GraphEffectToggle }
   | { type: "setEffect"; effect: GraphEffectToggle; enabled: boolean }
   | { type: "togglePanel"; panelId: string }
@@ -77,6 +80,7 @@ export interface GraphPluginContext {
   getEffectsState: () => GraphEffectsState;
   getDiagnosticsSnapshot: () => GraphDiagnosticsSnapshot | null;
   getAnalyticsSnapshot: () => GraphAnalyticsSnapshot | null;
+  getDisplayState: () => GraphDisplayStateSnapshot;
   isPanelOpen: (panelId: string) => boolean;
   dispatchAction: (action: GraphPluginActionRequest) => void;
 }

explorer/src/workspaces/GraphWorkspace/scene.ts

@@ -5,6 +5,8 @@ import { graph, type EdgeAttributes, type NodeAttributes } from "../../store/gra
 import type {
   GraphAnalyticsSnapshot,
   GraphCameraState,
+  GraphDisplayMeta,
+  GraphDisplayStateSnapshot,
   GraphDiagnosticsSnapshot,
   GraphEffectsState,
   GraphInteractionState,
@@ -22,6 +24,8 @@ export interface GraphSceneRuntime {
   scene: unknown;
   graph: GraphSceneGraph;
   displayGraph: GraphSceneGraph;
+  graphVersion: number;
+  layoutMode?: GraphDisplayMeta["layoutMode"];
   requestRender: () => void;
   getCameraState: () => GraphCameraState | null;
 }
@@ -37,7 +41,13 @@ export interface GraphSceneEventMap {
 }
 
 export interface GraphSceneProps extends GraphSceneEventMap {
+  graphVersion: number;
+  graphReady: boolean;
+  displayGraph: GraphSceneGraph;
+  displayMeta: GraphDisplayMeta;
+  displayState?: GraphDisplayStateSnapshot;
   selectedNodeId: string;
+  focusedNodeId: string;
   selectedEdgeId: string;
   activePath?: string[];
   activePathEdgeIds?: string[];
@@ -56,6 +66,8 @@ export interface GraphSceneProps extends GraphSceneEventMap {
 export interface GraphSceneHandle {
   fitView: () => void;
   focusNode: (nodeId: string) => void;
+  zoomIn: () => void;
+  zoomOut: () => void;
   getRuntime: () => GraphSceneRuntime | null;
   setLayoutRunning?: (running: boolean) => void;
 }

explorer/src/workspaces/GraphWorkspace/types.ts

@@ -1,4 +1,4 @@
-export type GraphViewMode = "focused" | "full";
+export type GraphViewMode = "focused" | "full" | "grouped";
 export type GraphLayoutSource = "provided" | "carried" | "runtime";
 export type GraphLayoutState = "idle" | "bootstrapping" | "running" | "stabilized" | "interactive" | "failed";
 export type GraphLoadPhase =
@@ -12,6 +12,7 @@ export type GraphLoadPhase =
 export type GraphLoadProgressKind = "determinate" | "indeterminate";
 export type GraphNodeInteractionState = "default" | "hovered" | "selected" | "neighbor" | "path" | "inactive" | "muted";
 export type GraphEdgeInteractionState = "default" | "backbone" | "hovered" | "selected" | "neighbor" | "path" | "inactive" | "muted";
+export type GraphSelectedNodeKind = "none" | "base" | "grouped" | "unavailable";
 
 export interface GraphCameraState {
   x: number;
@@ -31,6 +32,28 @@ export interface GraphInteractionState {
   isLayoutRunning: boolean;
 }
 
+export interface GraphDisplayStateSnapshot {
+  aggregationEnabled: boolean;
+  groupedViewAvailable: boolean;
+  groupedViewReason: string | null;
+  selectedRootNodeId: string | null;
+  selectedVisibleNeighborIds: string[];
+  selectedCollapsedNeighborIds: string[];
+  selectedNodeKind: GraphSelectedNodeKind;
+  canActivateFocused: boolean;
+  resolvedFocusedNodeId: string | null;
+  focusedUnavailableReason: string | null;
+}
+
+export type GraphDisplayLayoutMode = "base" | "mirrored" | "owned";
+
+export interface GraphDisplayMeta {
+  layoutMode: GraphDisplayLayoutMode;
+  positionSource: "store" | "display";
+  tracksStoreNodePositions: boolean;
+  hasSyntheticNodes: boolean;
+}
+
 export type GraphEffectToggle =
   | "pathPulseEnabled"
   | "pathFlowEnabled"
@@ -245,6 +268,10 @@ export interface GraphSelectedNodeState {
   valid_until?: string | null;
   properties: Record<string, unknown>;
   neighborCount: number;
+  visibleNeighborCount: number;
+  collapsedNeighborCount: number;
+  isNeighborhoodCollapsed: boolean;
+  canCollapseNeighborhood: boolean;
 }
 
 export interface GraphSelectedEdgeState {
@@ -260,6 +287,12 @@ export interface GraphSelectedEdgeState {
   provenanceCount: number;
   familySize: number;
   siblingCount: number;
+  isAggregated: boolean;
+  aggregateCount: number;
+  rawEdgeIds: string[];
+  bundleKind: "parallel" | "bidirectional" | "community" | null;
+  dominantEdgeType: string | null;
+  representativeWeight: number;
 }
 
 export interface GraphStageHandle {

explorer/tests/graphSceneState.display.test.ts

@@ -0,0 +1,259 @@
+import test from "node:test";
+import assert from "node:assert/strict";
+
+import {
+  batchMergeEdges,
+  batchMergeNodes,
+  clearGraph,
+} from "../src/store/graphStore.ts";
+import {
+  checkGroupedViewAvailability,
+  resolveDisplayGraph,
+  resolveGroupedDisplayNodeId,
+  resolveGroupedDisplayStateSnapshot,
+} from "../src/workspaces/GraphWorkspace/graphSceneState.ts";
+
+function addNode(id: string, semanticGroup = "entity") {
+  batchMergeNodes([
+    {
+      id,
+      attributes: {
+        label: id,
+        content: id,
+        x: 0,
+        y: 0,
+        size: 8,
+        color: "#63E6FF",
+        baseColor: "#63E6FF",
+        nodeType: semanticGroup,
+        semanticGroup,
+        properties: {},
+      },
+    },
+  ]);
+}
+
+function addEdge(id: string, source: string, target: string, weight = 1) {
+  batchMergeEdges([
+    {
+      id,
+      source,
+      target,
+      attributes: {
+        edgeType: "related_to",
+        weight,
+        properties: {},
+      },
+    },
+  ]);
+}
+
+test.beforeEach(() => {
+  clearGraph();
+});
+
+test.after(() => {
+  clearGraph();
+});
+
+test("resolveDisplayGraph bundles parallel edges in full view", () => {
+  addNode("a");
+  addNode("b");
+  addEdge("e1", "a", "b", 1);
+  addEdge("e2", "a", "b", 2);
+
+  const { graph } = resolveDisplayGraph("", [], [], "full", { aggregationEnabled: true });
+  assert.equal(graph.size, 1);
+
+  const edgeId = graph.edges()[0];
+  const attrs = graph.getEdgeAttributes(edgeId) as {
+    isAggregated?: boolean;
+    aggregateCount?: number;
+    rawEdgeIds?: string[];
+    bundleKind?: string;
+  };
+
+  assert.equal(attrs.isAggregated, true);
+  assert.equal(attrs.aggregateCount, 2);
+  assert.deepEqual(new Set(attrs.rawEdgeIds ?? []), new Set(["e1", "e2"]));
+  assert.equal(attrs.bundleKind, "parallel");
+});
+
+test("resolveDisplayGraph collapse keeps path neighbor visible", () => {
+  addNode("center");
+  for (let index = 0; index < 10; index += 1) {
+    const neighbor = `n${index}`;
+    addNode(neighbor);
+    addEdge(`edge-${index}`, "center", neighbor, 1);
+  }
+
+  const { state } = resolveDisplayGraph("center", ["center", "n9"], [], "full", {
+    aggregationEnabled: false,
+    collapsedNeighborhoodNodeIds: ["center"],
+  });
+
+  assert.equal(state.selectedRootNodeId, "center");
+  assert.equal(state.selectedVisibleNeighborIds.includes("n9"), true);
+  assert.equal(state.selectedVisibleNeighborIds.length, 9);
+  assert.equal(state.selectedCollapsedNeighborIds.length, 1);
+});
+
+test("resolveDisplayGraph grouped view emits community nodes and edges", () => {
+  const left = ["a1", "a2", "a3", "a4"];
+  const right = ["b1", "b2", "b3", "b4"];
+
+  [...left, ...right].forEach((nodeId, index) => {
+    addNode(nodeId, index < left.length ? "left" : "right");
+  });
+
+  let edgeIndex = 0;
+  for (let i = 0; i < left.length; i += 1) {
+    for (let j = 0; j < left.length; j += 1) {
+      if (i !== j) {
+        addEdge(`l-${edgeIndex++}`, left[i], left[j], 3);
+      }
+    }
+  }
+
+  for (let i = 0; i < right.length; i += 1) {
+    for (let j = 0; j < right.length; j += 1) {
+      if (i !== j) {
+        addEdge(`r-${edgeIndex++}`, right[i], right[j], 3);
+      }
+    }
+  }
+
+  addEdge("bridge-1", "a1", "b1", 0.1);
+  addEdge("bridge-2", "a2", "b2", 0.1);
+
+  const { graph, state } = resolveDisplayGraph("", [], [], "grouped", { aggregationEnabled: true });
+
+  assert.equal(state.groupedViewAvailable, true);
+
+  const communityNodes = graph.nodes().filter((nodeId) => nodeId.startsWith("__community__"));
+  assert.ok(communityNodes.length >= 2);
+
+  const hasCommunityEdge = graph
+    .edges()
+    .map((edgeId) => graph.getEdgeAttributes(edgeId) as { bundleKind?: string; isAggregated?: boolean; aggregateCount?: number })
+    .some((attrs) => attrs.bundleKind === "community" && attrs.isAggregated === true && Number(attrs.aggregateCount ?? 0) > 0);
+
+  assert.equal(hasCommunityEdge, true);
+});
+
+// ── resolveGroupedDisplayNodeId ──────────────────────────────────────────────
+
+test("resolveGroupedDisplayNodeId returns null for empty nodeId", () => {
+  const { graph: displayGraph } = resolveDisplayGraph("", [], [], "full", { aggregationEnabled: false });
+  assert.equal(resolveGroupedDisplayNodeId(displayGraph, ""), null);
+});
+
+test("resolveGroupedDisplayNodeId returns nodeId when it exists directly in display graph", () => {
+  addNode("x");
+  const { graph: displayGraph } = resolveDisplayGraph("", [], [], "full", { aggregationEnabled: false });
+  assert.equal(resolveGroupedDisplayNodeId(displayGraph, "x"), "x");
+});
+
+test("resolveGroupedDisplayNodeId resolves base node to its community node", () => {
+  const left = ["a1", "a2", "a3", "a4"];
+  const right = ["b1", "b2", "b3", "b4"];
+  [...left, ...right].forEach((nodeId, index) => addNode(nodeId, index < left.length ? "left" : "right"));
+
+  let edgeIndex = 0;
+  for (let i = 0; i < left.length; i += 1) {
+    for (let j = 0; j < left.length; j += 1) {
+      if (i !== j) addEdge(`l-${edgeIndex++}`, left[i], left[j], 3);
+    }
+  }
+  for (let i = 0; i < right.length; i += 1) {
+    for (let j = 0; j < right.length; j += 1) {
+      if (i !== j) addEdge(`r-${edgeIndex++}`, right[i], right[j], 3);
+    }
+  }
+  addEdge("bridge-1", "a1", "b1", 0.1);
+
+  const { graph: displayGraph } = resolveDisplayGraph("", [], [], "grouped", { aggregationEnabled: true });
+  const communityNodes = displayGraph.nodes().filter((n) => n.startsWith("__community__"));
+  assert.ok(communityNodes.length >= 2, "expected community nodes");
+
+  const resolved = resolveGroupedDisplayNodeId(displayGraph, "a1");
+  assert.ok(resolved !== null, "should resolve a1 to a community node");
+  assert.ok(resolved!.startsWith("__community__"), "resolved id should be a community node");
+});
+
+// ── resolveGroupedDisplayStateSnapshot ──────────────────────────────────────
+
+test("resolveGroupedDisplayStateSnapshot returns none-kind when no node selected", () => {
+  addNode("p");
+  addNode("q");
+  addEdge("e1", "p", "q");
+  const { graph: displayGraph } = resolveDisplayGraph("", [], [], "full", { aggregationEnabled: false });
+  const state = resolveGroupedDisplayStateSnapshot(displayGraph, "", {
+    groupedViewAvailable: true,
+    groupedViewReason: null,
+  });
+  assert.equal(state.selectedNodeKind, "none");
+  assert.equal(state.selectedRootNodeId, null);
+});
+
+test("resolveGroupedDisplayStateSnapshot maps selected base node to community in grouped graph", () => {
+  const left = ["c1", "c2", "c3", "c4"];
+  const right = ["d1", "d2", "d3", "d4"];
+  [...left, ...right].forEach((nodeId, index) => addNode(nodeId, index < left.length ? "left" : "right"));
+
+  let edgeIndex = 0;
+  for (let i = 0; i < left.length; i += 1) {
+    for (let j = 0; j < left.length; j += 1) {
+      if (i !== j) addEdge(`lc-${edgeIndex++}`, left[i], left[j], 3);
+    }
+  }
+  for (let i = 0; i < right.length; i += 1) {
+    for (let j = 0; j < right.length; j += 1) {
+      if (i !== j) addEdge(`rc-${edgeIndex++}`, right[i], right[j], 3);
+    }
+  }
+  addEdge("bridge-c1", "c1", "d1", 0.1);
+  addEdge("bridge-c2", "c2", "d2", 0.1);
+
+  const { graph: displayGraph } = resolveDisplayGraph("", [], [], "grouped", { aggregationEnabled: true });
+  const state = resolveGroupedDisplayStateSnapshot(displayGraph, "c1", {
+    groupedViewAvailable: true,
+    groupedViewReason: null,
+    selectedNodeKind: "grouped",
+  });
+
+  assert.ok(state.selectedRootNodeId !== null, "should resolve to a community node");
+  assert.ok(state.selectedRootNodeId!.startsWith("__community__"), "root should be a community node");
+  assert.equal(state.groupedViewAvailable, true);
+});
+
+// ── checkGroupedViewAvailability ─────────────────────────────────────────────
+
+test("checkGroupedViewAvailability returns unavailable on empty graph", () => {
+  const result = checkGroupedViewAvailability();
+  assert.equal(result.available, false);
+  assert.ok(typeof result.reason === "string" && result.reason.length > 0);
+});
+
+test("checkGroupedViewAvailability returns available when communities exist", () => {
+  const left = ["e1", "e2", "e3", "e4"];
+  const right = ["f1", "f2", "f3", "f4"];
+  [...left, ...right].forEach((nodeId, index) => addNode(nodeId, index < left.length ? "left" : "right"));
+
+  let edgeIndex = 0;
+  for (let i = 0; i < left.length; i += 1) {
+    for (let j = 0; j < left.length; j += 1) {
+      if (i !== j) addEdge(`le-${edgeIndex++}`, left[i], left[j], 3);
+    }
+  }
+  for (let i = 0; i < right.length; i += 1) {
+    for (let j = 0; j < right.length; j += 1) {
+      if (i !== j) addEdge(`re-${edgeIndex++}`, right[i], right[j], 3);
+    }
+  }
+  addEdge("bridge-e1", "e1", "f1", 0.1);
+
+  const result = checkGroupedViewAvailability();
+  assert.equal(result.available, true);
+  assert.equal(result.reason, null);
+});

plugins/.claude-plugin/README.md

@@ -1,195 +1,172 @@
 # Semantica Plugins (Community Guide)
 
-Semantica ships a shared plugin bundle under `plugins/` with skills, agents, and hooks for knowledge graphs, context graphs, decision intelligence, reasoning, explainability, provenance, ontology, and export workflows.
+> **v0.4.0** — 17 domain skills · 3 agents · 8 platform plugins · Knowledge Explorer UI
 
-This README covers installation across every supported platform.
+Semantica ships a shared plugin bundle under `plugins/` that works across every major AI coding assistant. Connect any supported platform to Semantica's knowledge graph engine for semantic extraction, decision intelligence, reasoning, provenance, ontology, and export workflows.
 
-## Supported Platforms
+---
 
-| Platform | Method | Config file |
-|---|---|---|
-| Claude Code | Native plugin bundle | `plugins/.claude-plugin/plugin.json` |
-| Cursor | Native plugin bundle | `plugins/.cursor-plugin/plugin.json` |
-| Codex CLI | Native plugin bundle | `plugins/.codex-plugin/plugin.json` |
-| Windsurf | MCP server + plugin bundle | `plugins/.windsurf-plugin/plugin.json` |
-| Cline (VS Code) | MCP server + plugin bundle | `plugins/.cline-plugin/plugin.json` |
-| Continue | MCP server | `plugins/.continue-plugin/plugin.json` |
-| VS Code | MCP server | `plugins/.vscode-plugin/plugin.json` |
-| Claude Desktop | MCP server | — (see MCP section below) |
-| Any MCP client | MCP server | `python -m semantica.mcp_server` |
+## Platform Plugins
+
+Semantica provides a dedicated plugin for each platform. Every plugin shares the same `skills/`, `agents/`, and `hooks/` bundle — only the manifest format differs.
+
+| # | Platform | Plugin Folder | Setup |
+|---|----------|--------------|-------|
+| 1 | **Claude Code** | `.claude-plugin/` | `claude --plugin-dir ./plugins` |
+| 2 | **Cursor** | `.cursor-plugin/` | Cursor Marketplace → refresh |
+| 3 | **Codex** | `.codex-plugin/` | Marketplace UI → install |
+| 4 | **Cline** | `.cline-plugin/` | Cline MCP settings |
+| 5 | **Windsurf** | `.windsurf-plugin/` | `mcp_config.json` |
+| 6 | **Continue** | `.continue-plugin/` | `~/.continue/config.json` |
+| 7 | **OpenClaw** | `.openclaw-plugin/` | `mcporter.json` |
+| 8 | **VS Code** | `.vscode-plugin/` | `settings.json` MCP entry |
+
+---
+
+## What's Included
+
+```
+plugins/
+├── skills/              # 17 domain skills (slash commands)
+├── agents/              # 3 specialized agents
+├── hooks/               # hooks.json
+├── .claude-plugin/      # Claude Code manifest + marketplace
+├── .cursor-plugin/      # Cursor manifest + marketplace
+├── .codex-plugin/       # Codex manifest + marketplace
+├── .cline-plugin/       # Cline manifest + marketplace
+├── .windsurf-plugin/    # Windsurf manifest + marketplace
+├── .continue-plugin/    # Continue manifest + marketplace
+├── .openclaw-plugin/    # OpenClaw manifest + marketplace
+└── .vscode-plugin/      # VS Code manifest + marketplace
+```
+
+### Skills (17)
+
+`extract` · `ingest` · `query` · `ontology` · `validate` · `deduplicate` · `embed` · `reason` · `decision` · `causal` · `temporal` · `provenance` · `policy` · `explain` · `export` · `change` · `visualize`
+
+### Agents (3)
+
+`decision-advisor` · `explainability` · `kg-assistant`
+
+---
 
 ## Prerequisites
 
-1. Clone the repository:
-
 ```bash
 git clone https://github.com/Hawksight-AI/semantica.git
 cd semantica
+pip install semantica        # Python 3.10+
 ```
 
-2. Ensure the plugin bundle exists at:
+---
 
-```text
-plugins/
-  skills/          ← 17 domain skills
-  agents/          ← 3 specialized agents
-  hooks/           ← hooks.json
-  .claude-plugin/  ← Claude Code manifest
-  .cursor-plugin/  ← Cursor manifest
-  .codex-plugin/   ← Codex CLI manifest
-  .windsurf-plugin/← Windsurf manifest + MCP config
-  .cline-plugin/   ← Cline manifest + MCP config
-  .continue-plugin/← Continue manifest + MCP config
-  .vscode-plugin/  ← VS Code manifest + MCP config
+## Knowledge Explorer (v0.4.0)
+
+Launch the interactive graph dashboard:
+
+```bash
+semantica-explorer --graph my_graph.json --port 8000
 ```
 
-## Plugin Contents
+Open **http://localhost:5174** to explore:
 
-- `skills/`: 17 domain skills (`causal`, `decision`, `explain`, `reason`, `temporal`, etc.)
-- `agents/`: specialized agents (`decision-advisor`, `explainability`, `kg-assistant`)
-- `hooks/hooks.json`: plugin hook configuration
-- `.claude-plugin/plugin.json`: Claude manifest
-- `.cursor-plugin/plugin.json`: Cursor manifest
-- `.codex-plugin/plugin.json`: Codex manifest
-- `*/marketplace.json`: local marketplace definitions
+- **Graph** — interactive canvas with ForceAtlas2 layout, path highlight, community coloring
+- **Decisions** — causal chains and outcome analysis
+- **Reasoning** — run deductive / abductive rules
+- **SPARQL** — Monaco editor for graph queries
+- **Vocabulary** — ontology concept tree
+- **Lineage** — provenance lineage diagram
+- **Import / Export** — JSON, RDF, Parquet, GraphML
 
-## Install and Use in Claude Code
+---
 
-### Local install (fastest)
+## Installation by Platform
 
-From the repository root:
+### Claude Code
 
 ```bash
 claude --plugin-dir ./plugins
 ```
 
-If your Claude setup uses plugin commands in-session, use:
+Or inside a session:
 
 ```bash
-/plugin install ./plugins
+/plugin marketplace add ./plugins
+/plugin install semantica@semantica-local
 ```
 
-### Install from a GitHub marketplace
+Verify:
 
-Add a marketplace hosted in git:
-
-```bash
-/plugin marketplace add <owner>/semantica
 ```
-
-Install Semantica from that marketplace:
-
-```bash
-/plugin install semantica@<marketplace-name>
-```
-
-### Verify in Claude
-
-Run one of these in chat:
-
-```text
 /semantica:decision list
 /semantica:explain decision <decision_id>
 ```
 
-If the plugin is installed correctly, Claude should recognize the `/semantica:*` skills.
+---
 
-## Install and Use in Codex
+### Cursor
+
+Cursor reads `.cursor-plugin/plugin.json` and `.cursor-plugin/marketplace.json` automatically. Publish the `plugins/` directory and refresh in Cursor Marketplace to pick up updates.
+
+Verify:
+
+```
+/semantica:visualize topology
+/semantica:reason deductive "IF Person(x) THEN Mortal(x)"
+```
+
+---
+
+### Codex
 
 1. Ensure your repo marketplace exists at `.agents/plugins/marketplace.json`.
-2. Point the plugin entry `source.path` to `./plugins` (or your chosen plugin directory).
+2. Set `source.path` to `./plugins` in the plugin entry.
 3. Restart Codex and install from the marketplace UI.
 
-Codex manifest used by this bundle:
+Verify:
 
-- `.codex-plugin/plugin.json`
-
-### Verify in Codex
-
-After install, run a Semantica skill command in chat, for example:
-
-```text
+```
 /semantica:causal chain --subject <decision_id> --depth 3
 ```
 
-## Install and Use in Cursor
+---
 
-Cursor reads plugin metadata from:
+### Cline
 
-- `.cursor-plugin/plugin.json`
-- `.cursor-plugin/marketplace.json`
-
-If you maintain a team/community plugin repo, publish this `plugins/` directory and refresh/reinstall in Cursor Marketplace to pick up updates.
-
-### Verify in Cursor
-
-Try one of these commands:
-
-```text
-/semantica:reason deductive "IF Person(x) THEN Mortal(x)"
-/semantica:visualize topology
-```
-
-## First Commands to Try
-
-After installing on any platform, these are good smoke tests:
-
-1. `/semantica:decision record <category> "<scenario>" "<reasoning>" <outcome> <confidence>`
-2. `/semantica:decision list`
-3. `/semantica:causal chain --subject <decision_id> --depth 3`
-4. `/semantica:explain decision <decision_id>`
-5. `/semantica:validate graph`
-
-## MCP Server (Windsurf · Cline · Continue · VS Code · Claude Desktop · Any tool)
-
-Semantica includes a full MCP server (`semantica/mcp_server.py`) that exposes 12 tools and 3 resources over stdio — compatible with any MCP-aware tool.
-
-### Start the server
-
-```bash
-python -m semantica.mcp_server
-```
-
-### Configure in your tool
-
-**Claude Desktop** — `~/Library/Application Support/Claude/claude_desktop_config.json` (macOS) or `%APPDATA%\Claude\claude_desktop_config.json` (Windows):
-
-```json
-{
-  "mcpServers": {
-    "semantica": {
-      "command": "python",
-      "args": ["-m", "semantica.mcp_server"]
-    }
-  }
-}
-```
-
-**Windsurf** — `~/.codeium/windsurf/mcp_config.json`:
-
-```json
-{
-  "mcpServers": {
-    "semantica": {
-      "command": "python",
-      "args": ["-m", "semantica.mcp_server"]
-    }
-  }
-}
-```
-
-**Cline** — Cline MCP settings panel → Add server:
+In Cline MCP settings, add:
 
 ```json
 {
   "semantica": {
     "command": "python",
-    "args": ["-m", "semantica.mcp_server"]
+    "args": ["-m", "semantica.mcp_server"],
+    "env": {}
   }
 }
 ```
 
-**Continue** — `~/.continue/config.json`:
+---
+
+### Windsurf
+
+Add to `~/.codeium/windsurf/mcp_config.json`:
+
+```json
+{
+  "mcpServers": {
+    "semantica": {
+      "command": "python",
+      "args": ["-m", "semantica.mcp_server"]
+    }
+  }
+}
+```
+
+---
+
+### Continue
+
+Add to `~/.continue/config.json`:
 
 ```json
 {
@@ -203,7 +180,50 @@ python -m semantica.mcp_server
 }
 ```
 
-**VS Code** — `settings.json`:
+All 17 Semantica skills appear in the `@semantica` context provider dropdown.
+
+---
+
+### OpenClaw
+
+Add to `~/.openclaw/mcporter.json`:
+
+```json
+{
+  "mcpServers": {
+    "semantica": {
+      "command": "python",
+      "args": ["-m", "semantica.mcp_server"],
+      "transport": "stdio"
+    }
+  }
+}
+```
+
+Then restart the gateway:
+
+```bash
+openclaw gateway restart
+```
+
+---
+
+### VS Code
+
+Add to `settings.json` (GitHub Copilot Chat):
+
+```json
+{
+  "github.copilot.chat.mcp.servers": {
+    "semantica": {
+      "command": "python",
+      "args": ["-m", "semantica.mcp_server"]
+    }
+  }
+}
+```
+
+Or for the VS Code MCP extension:
 
 ```json
 {
@@ -216,40 +236,25 @@ python -m semantica.mcp_server
 }
 ```
 
-### Available MCP tools
+---
 
-| Tool | Description |
-|---|---|
-| `extract_entities` | Named entity recognition from text |
-| `extract_relations` | Relation and triplet extraction from text |
-| `record_decision` | Record a decision with full context and metadata |
-| `query_decisions` | Query recorded decisions by natural language or category |
-| `find_precedents` | Find past decisions similar to a scenario |
-| `get_causal_chain` | Trace upstream/downstream causal chain from a decision |
-| `add_entity` | Add a node/entity to the knowledge graph |
-| `add_relationship` | Add a directed edge between two entities |
-| `run_reasoning` | Run IF/THEN rules over facts to derive new facts |
-| `get_graph_analytics` | PageRank centrality and community detection |
-| `export_graph` | Export graph as Turtle, JSON-LD, N-Triples, or JSON |
-| `get_graph_summary` | Node count, decision count, graph status |
+## First Commands to Try
 
-### Available MCP resources
+After installing on any platform:
 
-| URI | Description |
-|---|---|
-| `semantica://graph/summary` | High-level graph statistics |
-| `semantica://decisions/list` | All recorded decisions |
-| `semantica://schema/info` | Server info and capability list |
+```
+/semantica:decision record <category> "<scenario>" "<reasoning>" <outcome> <confidence>
+/semantica:decision list
+/semantica:causal chain --subject <decision_id> --depth 3
+/semantica:explain decision <decision_id>
+/semantica:validate graph
+/semantica:visualize topology
+```
 
-### Environment variables
-
-| Variable | Description |
-|---|---|
-| `SEMANTICA_KG_PATH` | Path to a persisted graph to load on start |
-| `SEMANTICA_LOG_LEVEL` | Log level: DEBUG, INFO, WARNING (default: WARNING) |
+---
 
 ## Community Notes
 
-- Keep plugin name/version/keywords updated in each manifest before publishing.
-- Keep skill frontmatter consistent (`name` + `description`) for reliable discovery.
-- For open-source sharing, include this folder as-is so skills, agents, and hooks remain bundled.
+- Keep `name` / `version` / `keywords` updated in each manifest before publishing.
+- Keep skill frontmatter (`name` + `description`) consistent for reliable discovery.
+- Include `plugins/` as-is when sharing — skills, agents, and hooks must stay bundled.

plugins/.claude-plugin/marketplace.json

@@ -1,5 +1,9 @@
 {
   "name": "semantica-local",
+  "owner": {
+    "name": "Hawksight AI",
+    "url": "https://github.com/Hawksight-AI/semantica"
+  },
   "plugins": [
     {
       "name": "semantica",

plugins/.claude-plugin/plugin.json

@@ -25,6 +25,5 @@
     "mcp"
   ],
   "skills": "./skills",
-  "agents": "./agents",
-  "hooks": "./hooks/hooks.json"
+  "agents": "./agents"
 }

plugins/.cline-plugin/README.md

@@ -1,6 +1,6 @@
 # Semantica — Cline Plugin
 
-Adds all 17 Semantica skills, 3 agents, and hook configuration to Cline (VS Code extension).
+> **v0.4.0** — Adds all 17 Semantica skills, 3 agents, and hook configuration to Cline (VS Code extension).
 
 ## MCP Server Setup (recommended)
 
@@ -16,9 +16,17 @@ In Cline settings, add a new MCP server:
 }
 ```
 
-Cline will discover all 12 Semantica tools automatically on connection.
+Cline will discover all 17 Semantica skills and 3 agents automatically on connection.
+
+## Knowledge Explorer
+
+```bash
+semantica-explorer --graph my_graph.json --port 8000
+```
+
+Open `http://localhost:5174` for the interactive dashboard.
 
 ## Requirements
 
-- Python 3.8+
+- Python 3.10+
 - `pip install semantica`

plugins/.continue-plugin/README.md

@@ -1,6 +1,6 @@
 # Semantica — Continue Plugin
 
-Adds Semantica as an MCP server and context provider to [Continue.dev](https://continue.dev).
+> **v0.4.0** — Adds Semantica as an MCP server and context provider to [Continue.dev](https://continue.dev).
 
 ## MCP Server Setup
 
@@ -18,9 +18,17 @@ Add to `~/.continue/config.json`:
 }
 ```
 
-Continue will show all Semantica tools in the `@semantica` context provider dropdown.
+Continue will show all 17 Semantica skills in the `@semantica` context provider dropdown.
+
+## Knowledge Explorer
+
+```bash
+semantica-explorer --graph my_graph.json --port 8000
+```
+
+Open `http://localhost:5174` for the interactive graph dashboard.
 
 ## Requirements
 
-- Python 3.8+
+- Python 3.10+
 - `pip install semantica`

plugins/.openclaw-plugin/README.md

@@ -1,6 +1,6 @@
 # Semantica — OpenClaw Plugin
 
-Adds all 17 Semantica skills, 3 agents, and the full MCP integration to [OpenClaw](https://openclaw.ai) — the open-source personal AI agent platform.
+> **v0.4.0** — Adds all 17 Semantica skills, 3 agents, and the full MCP integration to [OpenClaw](https://openclaw.ai) — the open-source personal AI agent platform.
 
 ## MCP Server Setup (recommended)
 
@@ -32,7 +32,7 @@ Paste the following into your OpenClaw `mcporter.json` (usually `~/.openclaw/mcp
 openclaw gateway restart
 ```
 
-OpenClaw will automatically discover all 12 Semantica tools and 3 resources.
+OpenClaw will automatically discover all 17 Semantica tools and 3 agents.
 
 ## Skills
 
@@ -57,6 +57,6 @@ See [`integrations/openclaw/README.md`](../../integrations/openclaw/README.md) f
 
 ## Requirements
 
-- Python 3.8+
+- Python 3.10+
 - `pip install semantica`
 - OpenClaw — [openclaw.ai](https://openclaw.ai)

plugins/.vscode-plugin/README.md

@@ -1,6 +1,6 @@
 # Semantica — VS Code Plugin
 
-Adds Semantica as an MCP server to VS Code (via GitHub Copilot Chat or any MCP-aware extension).
+> **v0.4.0** — Adds Semantica as an MCP server to VS Code (via GitHub Copilot Chat or any MCP-aware extension).
 
 ## MCP Server Setup
 
@@ -30,7 +30,19 @@ Or if using the VS Code MCP extension directly:
 }
 ```
 
+VS Code will discover all 17 Semantica skills and 3 agents automatically on connection.
+
+## Knowledge Explorer
+
+Launch the interactive graph dashboard from the terminal:
+
+```bash
+semantica-explorer --graph my_graph.json --port 8000
+```
+
+Open `http://localhost:5174` to explore nodes, edges, decisions, SPARQL, lineage, and more.
+
 ## Requirements
 
-- Python 3.8+
+- Python 3.10+
 - `pip install semantica`

plugins/.windsurf-plugin/README.md

@@ -1,6 +1,6 @@
 # Semantica — Windsurf Plugin
 
-Adds all 17 Semantica skills, 3 agents, and hook configuration to Windsurf.
+> **v0.4.0** — Adds all 17 Semantica skills, 3 agents, and hook configuration to Windsurf.
 
 ## MCP Server Setup (recommended)
 
@@ -17,13 +17,23 @@ Add to your Windsurf MCP config (`~/.codeium/windsurf/mcp_config.json`):
 }
 ```
 
-Windsurf will then have access to all 12 Semantica tools (extract, record_decision, query_decisions, find_precedents, get_causal_chain, add_entity, add_relationship, run_reasoning, get_graph_analytics, export_graph, and more) directly in the AI panel.
+Windsurf will have access to all 17 Semantica skills (`extract`, `record_decision`, `query_decisions`, `find_precedents`, `get_causal_chain`, `add_entity`, `add_relationship`, `run_reasoning`, `get_graph_analytics`, `export_graph`, and more) directly in the AI panel.
 
 ## Skills
 
-All 17 skills under `plugins/skills/` are available as slash commands once the plugin is loaded.
+All 17 skills under `plugins/skills/` are available as slash commands once the plugin is loaded:
+
+`extract` · `ingest` · `query` · `ontology` · `validate` · `deduplicate` · `embed` · `reason` · `decision` · `causal` · `temporal` · `provenance` · `policy` · `explain` · `export` · `change` · `visualize`
+
+## Knowledge Explorer
+
+```bash
+semantica-explorer --graph my_graph.json --port 8000
+```
+
+Open `http://localhost:5174` for the interactive graph dashboard.
 
 ## Requirements
 
-- Python 3.8+
+- Python 3.10+
 - `pip install semantica`

pyproject.toml

@@ -90,7 +90,7 @@ llm-groq = ["groq>=0.4.0"]
 llm-gemini = ["google-genai>=0.1.0"]
 llm-anthropic = ["anthropic>=0.18.0"]
 llm-ollama = ["ollama>=0.1.0"]
-llm-deepseek = ["deepseek>=0.1.0"]
+llm-deepseek = ["openai>=1.0.0"]
 llm-litellm = ["litellm>=1.0.0"]
 llm-instructor = ["instructor>=1.0.0"]
 

semantica/__init__.py

@@ -10,7 +10,7 @@ Main exports:
     - Config: Configuration management
 """
 
-__version__ = "0.3.0"
+__version__ = "0.4.0"
 __author__ = "Semantica Contributors"
 __license__ = "MIT"
 

semantica/explorer/app.py

@@ -10,7 +10,7 @@ from typing import Optional
 
 from fastapi import FastAPI, HTTPException, Request, WebSocket, WebSocketDisconnect
 from fastapi.middleware.cors import CORSMiddleware
-from fastapi.responses import FileResponse, JSONResponse
+from fastapi.responses import FileResponse, HTMLResponse, JSONResponse
 from fastapi.staticfiles import StaticFiles
 
 from .. import __version__
@@ -19,7 +19,12 @@ from .ws import ConnectionManager
 
 
 def _install_mutation_bridge(app: FastAPI, session: GraphSession) -> None:
+    previous_callback = getattr(session.graph, "mutation_callback", None)
+
     def on_mutation(event_type: str, entity_id: str, payload: dict) -> None:
+        session.handle_graph_mutation(event_type, entity_id, payload)
+        if callable(previous_callback):
+            previous_callback(event_type, entity_id, payload)
         loop = getattr(app.state, "event_loop", None)
         manager = getattr(app.state, "ws_manager", None)
         if loop is None or manager is None or loop.is_closed():
@@ -127,6 +132,17 @@ def create_app(session: Optional[GraphSession] = None) -> FastAPI:
         except WebSocketDisconnect:
             manager.disconnect(websocket)
 
+    @app.get("/", include_in_schema=False)
+    async def root():
+        index_path = Path(__file__).resolve().parent.parent / "static" / "index.html"
+        if index_path.is_file():
+            return FileResponse(index_path)
+        return HTMLResponse(
+            '<!doctype html><html lang="en"><head><meta charset="UTF-8">'
+            '<title>Semantica Knowledge Explorer</title></head>'
+            '<body><div id="root"></div></body></html>'
+        )
+
     @app.get("/api/health")
     async def health():
         return {"status": "healthy"}
@@ -139,6 +155,17 @@ def create_app(session: Optional[GraphSession] = None) -> FastAPI:
             "status": "active",
         }
 
+    @app.get("/", include_in_schema=False)
+    async def root():
+        index_path = Path(__file__).resolve().parent.parent / "static" / "index.html"
+        if index_path.is_file():
+            return FileResponse(index_path)
+        return HTMLResponse(
+            '<!doctype html><html lang="en"><head><meta charset="UTF-8">'
+            '<title>Semantica Knowledge Explorer</title></head>'
+            '<body><div id="root"></div></body></html>'
+        )
+
     static_dir = Path(__file__).resolve().parent.parent / "static"
     if static_dir.is_dir():
         assets_dir = static_dir / "assets"

semantica/explorer/routes/annotations.py

@@ -8,7 +8,7 @@ modify Semantica core.
 import asyncio
 from typing import Optional
 
-from fastapi import APIRouter, Depends, Query
+from fastapi import APIRouter, Depends, HTTPException, Query
 
 from ..dependencies import get_session
 from ..schemas import AnnotationCreate, AnnotationResponse
@@ -35,7 +35,7 @@ async def create_annotation(
     """Create a new annotation on a node."""
     node = await asyncio.to_thread(session.get_node, body.node_id)
     if node is None:
-        raise KeyError(body.node_id)
+        raise HTTPException(status_code=404, detail=f"Node '{body.node_id}' not found")
 
     ann_data = body.model_dump()
     ann_id = await asyncio.to_thread(session.add_annotation, ann_data)
@@ -61,5 +61,5 @@ async def delete_annotation(
     """Delete an annotation by ID."""
     deleted = await asyncio.to_thread(session.delete_annotation, annotation_id)
     if not deleted:
-        raise KeyError(annotation_id)
+        raise HTTPException(status_code=404, detail=f"Annotation '{annotation_id}' not found")
     return None

semantica/explorer/routes/decisions.py

@@ -5,7 +5,7 @@ Decision routes using ContextGraph-native fallbacks.
 import asyncio
 from typing import Optional
 
-from fastapi import APIRouter, Depends, Query
+from fastapi import APIRouter, Depends, HTTPException, Query
 
 from ..dependencies import get_session
 from ..schemas import CausalChainResponse, ComplianceResponse, DecisionResponse
@@ -59,7 +59,7 @@ async def get_decision(
 ):
     node = await asyncio.to_thread(session.get_node, decision_id)
     if node is None:
-        raise KeyError(decision_id)
+        raise HTTPException(status_code=404, detail=f"Decision '{decision_id}' not found")
     return _node_to_decision(node)
 
 
@@ -70,7 +70,7 @@ async def get_causal_chain(
 ):
     node = await asyncio.to_thread(session.get_node, decision_id)
     if node is None:
-        raise KeyError(decision_id)
+        raise HTTPException(status_code=404, detail=f"Decision '{decision_id}' not found")
 
     neighbors = await asyncio.to_thread(session.get_neighbors, decision_id, 5)
     chain = [
@@ -94,7 +94,7 @@ async def get_precedents(
 ):
     node = await asyncio.to_thread(session.get_node, decision_id)
     if node is None:
-        raise KeyError(decision_id)
+        raise HTTPException(status_code=404, detail=f"Decision '{decision_id}' not found")
 
     properties = node.get("properties", {})
     category = str(properties.get("category", ""))
@@ -132,7 +132,7 @@ async def check_compliance(
 ):
     node = await asyncio.to_thread(session.get_node, decision_id)
     if node is None:
-        raise KeyError(decision_id)
+        raise HTTPException(status_code=404, detail=f"Decision '{decision_id}' not found")
 
     edges, _ = await asyncio.to_thread(session.get_edges, skip=0, limit=999_999)
     violation_types = {"violates", "non_compliant", "breaches"}

semantica/explorer/routes/enrich.py

@@ -6,7 +6,7 @@ import asyncio
 import re
 from typing import Dict, List, Optional, Tuple
 
-from fastapi import APIRouter, Depends
+from fastapi import APIRouter, Depends, HTTPException
 
 from ..dependencies import get_session
 from ..schemas import (
@@ -136,11 +136,11 @@ def _apply_inferred_edges(
             continue
         source, target = args
         if session.get_node(source) is None:
-            session.graph.add_node(source, "entity", content=source)
+            session.add_node(source, "entity", content=source)
         if session.get_node(target) is None:
-            session.graph.add_node(target, "entity", content=target)
+            session.add_node(target, "entity", content=target)
         edge_type = body.inferred_edge_type or predicate
-        session.graph.add_edge(
+        session.add_edge(
             source,
             target,
             edge_type=edge_type,
@@ -173,11 +173,12 @@ async def extract_entities(
             relations=[_safe_dict(relation) for relation in rel_list],
         )
     except ImportError:
-        raise ValueError(
-            "semantic_extract module not available. Ensure spacy and transformers are installed."
+        raise HTTPException(
+            status_code=503,
+            detail="semantic_extract module not available. Ensure spacy and transformers are installed.",
         )
     except Exception as exc:
-        raise ValueError(f"Extraction failed: {exc}")
+        raise HTTPException(status_code=422, detail=f"Extraction failed: {exc}")
 
 
 @router.post("/api/enrich/links", response_model=LinkPredictionResponse)
@@ -187,11 +188,11 @@ async def predict_links(
 ):
     predictor = session.link_predictor
     if predictor is None:
-        raise ValueError("LinkPredictor not available; KG extras may not be installed.")
+        raise HTTPException(status_code=503, detail="LinkPredictor not available; KG extras may not be installed.")
 
     node = await asyncio.to_thread(session.get_node, body.node_id)
     if node is None:
-        raise KeyError(body.node_id)
+        raise HTTPException(status_code=404, detail=f"Node '{body.node_id}' not found")
 
     nodes, _ = await asyncio.to_thread(session.get_nodes, skip=0, limit=999_999)
     edges, _ = await asyncio.to_thread(session.get_edges, skip=0, limit=999_999)
@@ -248,9 +249,9 @@ async def detect_duplicates(
         duplicate_list = duplicates if isinstance(duplicates, list) else getattr(duplicates, "duplicates", [])
         return DedupResponse(duplicates=[_safe_dict(item) for item in duplicate_list], total_flagged=len(duplicate_list))
     except ImportError:
-        raise ValueError("Deduplication module not available.")
+        raise HTTPException(status_code=503, detail="Deduplication module not available.")
     except Exception as exc:
-        raise ValueError(f"Dedup scan failed: {exc}")
+        raise HTTPException(status_code=422, detail=f"Dedup scan failed: {exc}")
 
 
 @router.post("/api/reason", response_model=ReasoningResponse)
@@ -295,7 +296,7 @@ async def merge_nodes(
 
     node = await asyncio.to_thread(session.get_node, primary_id)
     if node is None:
-        raise ValueError(f"Primary node {primary_id} not found")
+        raise HTTPException(status_code=404, detail=f"Primary node '{primary_id}' not found")
 
     def _do_merge() -> tuple[list[str], int]:
         removed: list[str] = []
@@ -353,4 +354,6 @@ async def merge_nodes(
         return removed, edges_updated
 
     removed_ids, edges_updated = await asyncio.to_thread(_do_merge)
+    if removed_ids:
+        await asyncio.to_thread(session.rebuild_search_index)
     return MergeResponse(merged_into=primary_id, removed_ids=removed_ids, edges_updated=edges_updated)

semantica/explorer/routes/graph.py

@@ -6,8 +6,9 @@ import asyncio
 from enum import Enum
 from typing import Optional
 
-from fastapi import APIRouter, Depends, Query
+from fastapi import APIRouter, Depends, HTTPException, Query
 
+from ...utils.helpers import classify_path_distance
 from ..dependencies import get_session
 from ..schemas import (
     EdgeListResponse,
@@ -83,7 +84,7 @@ async def get_node(
 ):
     node = await asyncio.to_thread(session.get_node, node_id)
     if node is None:
-        raise KeyError(node_id)
+        raise HTTPException(status_code=404, detail=f"Node '{node_id}' not found")
     return _node_response(node)
 
 
@@ -141,16 +142,19 @@ class _PathAlgorithm(str, Enum):
     dijkstra = "dijkstra"
 
 
+
+
 @router.get("/node/{node_id}/path", response_model=PathResponse)
 async def find_path(
     node_id: str,
     target: str = Query(..., description="Target node ID"),
     algorithm: _PathAlgorithm = Query(_PathAlgorithm.bfs, description="Algorithm: bfs or dijkstra"),
+    directed: bool = Query(True, description="If false, treat edges as undirected for traversal"),
     session: GraphSession = Depends(get_session),
 ):
     path_finder = session.path_finder
     if path_finder is None:
-        raise ValueError("PathFinder not available; KG extras may not be installed.")
+        raise HTTPException(status_code=503, detail="PathFinder not available; KG extras may not be installed.")
 
     graph_dict = await asyncio.to_thread(session.build_graph_dict)
     path_fn = (
@@ -158,12 +162,19 @@ async def find_path(
         if algorithm == _PathAlgorithm.dijkstra
         else path_finder.bfs_shortest_path
     )
-    result = await asyncio.to_thread(path_fn, graph_dict, node_id, target)
+    try:
+        result = await asyncio.to_thread(path_fn, graph_dict, node_id, target, directed=directed)
+    except Exception as exc:
+        raise HTTPException(status_code=404, detail=f"No path found from '{node_id}' to '{target}': {exc}")
 
     path_nodes = result.get("path", []) if isinstance(result, dict) else (result or [])
+    if not path_nodes:
+        raise HTTPException(status_code=404, detail=f"No path found from '{node_id}' to '{target}'")
+
     total_weight = result.get("total_weight", 0.0) if isinstance(result, dict) else 0.0
     edge_ids = await asyncio.to_thread(session.resolve_path_edge_ids, path_nodes)
 
+    hop_count = len(path_nodes) - 1 if path_nodes else 0
     return PathResponse(
         source=node_id,
         target=target,
@@ -171,6 +182,9 @@ async def find_path(
         path=path_nodes,
         edge_ids=edge_ids,
         total_weight=total_weight,
+        directed=directed,
+        hop_count=hop_count,
+        distance_band=classify_path_distance(hop_count),
     )
 
 

semantica/explorer/routes/provenance.py

@@ -1,4 +1,4 @@
-"""
+"""
 Provenance routes for lineage visualization and exportable reports.
 """
 
@@ -9,33 +9,13 @@ from typing import Any, Dict, List, Optional
 import networkx as nx
 from fastapi import APIRouter, Depends, Query
 from fastapi.responses import PlainTextResponse, Response
-from pydantic import BaseModel
 
 from ..dependencies import get_session
+from ..schemas import ProvenanceEdge, ProvenanceNode, ProvenanceResponse
 from ..session import GraphSession
 
 router = APIRouter(prefix="/api/provenance", tags=["Power User Tools"])
 
-
-class ProvenanceNode(BaseModel):
-    id: str
-    label: str
-    prov_type: str
-    parent_id: str
-
-
-class ProvenanceEdge(BaseModel):
-    id: str
-    source: str
-    target: str
-    label: str
-
-
-class ProvenanceResponse(BaseModel):
-    nodes: List[ProvenanceNode]
-    edges: List[ProvenanceEdge]
-
-
 _AGENT_TYPES = {"person", "organization", "system", "agent"}
 _ACTIVITY_TYPES = {"action", "event", "process", "activity", "decision", "publication"}
 
@@ -67,7 +47,7 @@ def _build_provenance(session: GraphSession, node_id: Optional[str] = None) -> d
         if edge.source_id in hop_nodes or edge.target_id in hop_nodes:
             graph.add_edge(edge.source_id, edge.target_id, label=edge.edge_type)
 
-    subgraph = nx.ego_graph(graph, node_id, radius=2, undirected=False)
+    subgraph = nx.ego_graph(graph, node_id, radius=2, undirected=True)
     provenance_nodes: List[Dict[str, Any]] = []
     for graph_node_id in subgraph.nodes():
         node = session.graph.nodes.get(graph_node_id)
@@ -85,12 +65,19 @@ def _build_provenance(session: GraphSession, node_id: Optional[str] = None) -> d
 
     provenance_edges: List[Dict[str, Any]] = []
     for source, target, data in subgraph.edges(data=True):
+        if target == node_id:
+            direction = "upstream"
+        elif source == node_id:
+            direction = "downstream"
+        else:
+            direction = "lateral"
         provenance_edges.append(
             {
                 "id": f"{source}-{target}",
                 "source": source,
                 "target": target,
                 "label": data.get("label", "related_to"),
+                "direction": direction,
             }
         )
 
@@ -104,7 +91,7 @@ def _build_report(session: GraphSession, node_id: str) -> Dict[str, Any]:
         "node_id": node_id,
         "label": node.get("content", node_id) if node else node_id,
         "type": node.get("type", "entity") if node else "entity",
-        "properties": node.get("properties", {}) if node else {},
+        "properties": node.get("metadata", node.get("properties", {})) if node else {},
         "lineage": provenance,
     }
 
@@ -129,9 +116,29 @@ def _render_markdown(report: Dict[str, Any]) -> str:
     for node in report.get("lineage", {}).get("nodes", []):
         lines.append(f"- `{node['id']}` ({node['prov_type']}): {node['label']}")
 
-    lines.extend(["", "## Lineage Edges"])
-    for edge in report.get("lineage", {}).get("edges", []):
-        lines.append(f"- `{edge['source']}` -[{edge['label']}]-> `{edge['target']}`")
+    edges = report.get("lineage", {}).get("edges", [])
+    grouped_edges: Dict[str, List] = {"upstream": [], "downstream": [], "lateral": []}
+    for edge in edges:
+        direction = edge.get("direction", "lateral")
+        if direction not in grouped_edges:
+            direction = "lateral"
+        grouped_edges[direction].append(edge)
+
+    if grouped_edges["upstream"]:
+        lines.extend(["", "## Upstream"])
+        for edge in grouped_edges["upstream"]:
+            lines.append(f"- `{edge['source']}` -[{edge['label']}]-> `{edge['target']}`")
+
+    if grouped_edges["downstream"]:
+        lines.extend(["", "## Downstream"])
+        for edge in grouped_edges["downstream"]:
+            lines.append(f"- `{edge['source']}` -[{edge['label']}]-> `{edge['target']}`")
+
+    if grouped_edges["lateral"]:
+        lines.extend(["", "## Lateral"])
+        for edge in grouped_edges["lateral"]:
+            lines.append(f"- `{edge['source']}` -[{edge['label']}]-> `{edge['target']}`")
+
     return "\n".join(lines)
 
 

semantica/explorer/routes/temporal.py

@@ -103,7 +103,7 @@ async def temporal_patterns(
 
         detector = TemporalPatternDetector()
         graph_dict = await asyncio.to_thread(session.build_graph_dict)
-        patterns = await asyncio.to_thread(detector.detect_patterns, graph_dict)
+        patterns = await asyncio.to_thread(detector.detect_temporal_patterns, graph_dict)
         if isinstance(patterns, dict):
             patterns = patterns.get("patterns", [])
         return TemporalPatternResponse(patterns=patterns if isinstance(patterns, list) else [])

semantica/explorer/schemas.py

@@ -67,6 +67,9 @@ class PathResponse(BaseModel):
     path: List[str]
     edge_ids: List[str] = Field(default_factory=list)
     total_weight: float = 0.0
+    directed: bool = True
+    hop_count: int = 0
+    distance_band: str = "direct"
 
 
 class GraphStatsResponse(BaseModel):
@@ -285,3 +288,23 @@ class MergeResponse(BaseModel):
     merged_into: str
     removed_ids: List[str]
     edges_updated: int
+
+
+class ProvenanceNode(BaseModel):
+    id: str
+    label: str
+    prov_type: str
+    parent_id: Optional[str] = None
+
+
+class ProvenanceEdge(BaseModel):
+    id: str
+    source: str
+    target: str
+    label: str
+    direction: str
+
+
+class ProvenanceResponse(BaseModel):
+    nodes: List[ProvenanceNode]
+    edges: List[ProvenanceEdge]

semantica/explorer/search_index.py

@@ -0,0 +1,398 @@
+"""
+Explorer-local in-memory node search index.
+"""
+
+from __future__ import annotations
+
+import bisect
+import heapq
+import re
+from collections import OrderedDict, defaultdict
+from dataclasses import dataclass
+from typing import Any, DefaultDict, Dict, Iterable, List, Optional, Tuple
+
+_TOKEN_RE = re.compile(r"[a-z0-9]+")
+_WHITESPACE_RE = re.compile(r"\s+")
+_CURATED_ALIAS_KEYS = (
+    "label",
+    "name",
+    "title",
+    "pref_label",
+    "preferred_label",
+    "prefLabel",
+    "aliases",
+    "alias",
+    "synonyms",
+    "synonym",
+    "symbol",
+    "display_name",
+    "displayName",
+    "text",
+    "content",
+)
+
+
+def _normalize_text(value: Any) -> str:
+    if value is None:
+        return ""
+    text = str(value).strip().lower()
+    if not text:
+        return ""
+    return _WHITESPACE_RE.sub(" ", text)
+
+
+def _tokenize(text: str) -> Tuple[str, ...]:
+    if not text:
+        return ()
+    return tuple(dict.fromkeys(_TOKEN_RE.findall(text)))
+
+
+def _collect_text_fragments(value: Any, fragments: List[str], *, limit: int = 64) -> None:
+    if value is None or len(fragments) >= limit:
+        return
+    if isinstance(value, dict):
+        for nested in value.values():
+            _collect_text_fragments(nested, fragments, limit=limit)
+            if len(fragments) >= limit:
+                return
+        return
+    if isinstance(value, (list, tuple, set)):
+        for nested in value:
+            _collect_text_fragments(nested, fragments, limit=limit)
+            if len(fragments) >= limit:
+                return
+        return
+
+    normalized = _normalize_text(value)
+    if normalized:
+        fragments.append(normalized)
+
+
+def _coerce_float(value: Any) -> Optional[float]:
+    if value is None or value == "":
+        return None
+    try:
+        return float(value)
+    except (TypeError, ValueError):
+        return None
+
+
+@dataclass(frozen=True)
+class IndexedNodeDocument:
+    node_id: str
+    normalized_id: str
+    node_type: str
+    exact_terms: frozenset[str]
+    tokens: frozenset[str]
+    primary_text: str
+    secondary_text: str
+    confidence: Optional[float]
+    tags: Tuple[str, ...]
+
+
+class GraphSearchIndex:
+    def __init__(
+        self,
+        *,
+        cache_size: int = 128,
+        prefix_min_length: int = 2,
+        prefix_max_length: int = 12,
+        secondary_scan_limit: int = 12000,
+    ) -> None:
+        self.cache_size = cache_size
+        self.prefix_min_length = prefix_min_length
+        self.prefix_max_length = prefix_max_length
+        self.secondary_scan_limit = secondary_scan_limit
+        self._documents: Dict[str, IndexedNodeDocument] = {}
+        self._exact_index: DefaultDict[str, set[str]] = defaultdict(set)
+        self._token_index: DefaultDict[str, set[str]] = defaultdict(set)
+        self._prefix_index: DefaultDict[str, set[str]] = defaultdict(set)
+        self._ordered_node_ids: List[str] = []
+        self._cache: OrderedDict[Tuple[Any, ...], List[Tuple[str, float]]] = OrderedDict()
+
+    def rebuild(self, nodes: Iterable[Dict[str, Any]]) -> None:
+        self._documents.clear()
+        self._exact_index.clear()
+        self._token_index.clear()
+        self._prefix_index.clear()
+        self._ordered_node_ids = []
+        self.clear_cache()
+
+        for node in nodes:
+            self.upsert(node, clear_cache=False)
+
+        self._ordered_node_ids.sort()
+
+    def clear_cache(self) -> None:
+        self._cache.clear()
+
+    def remove(self, node_id: str, *, clear_cache: bool = True) -> None:
+        existing = self._documents.pop(node_id, None)
+        if existing is None:
+            return
+
+        for term in existing.exact_terms:
+            bucket = self._exact_index.get(term)
+            if bucket is None:
+                continue
+            bucket.discard(node_id)
+            if not bucket:
+                self._exact_index.pop(term, None)
+
+        for token in existing.tokens:
+            bucket = self._token_index.get(token)
+            if bucket is None:
+                continue
+            bucket.discard(node_id)
+            if not bucket:
+                self._token_index.pop(token, None)
+
+            for length in range(self.prefix_min_length, min(len(token), self.prefix_max_length) + 1):
+                prefix = token[:length]
+                prefix_bucket = self._prefix_index.get(prefix)
+                if prefix_bucket is None:
+                    continue
+                prefix_bucket.discard(node_id)
+                if not prefix_bucket:
+                    self._prefix_index.pop(prefix, None)
+
+        pos = bisect.bisect_left(self._ordered_node_ids, node_id)
+        if pos < len(self._ordered_node_ids) and self._ordered_node_ids[pos] == node_id:
+            self._ordered_node_ids.pop(pos)
+
+        if clear_cache:
+            self.clear_cache()
+
+    def upsert(self, node: Dict[str, Any], *, clear_cache: bool = True) -> None:
+        node_id = str(node.get("id", "")).strip()
+        if not node_id:
+            return
+
+        self.remove(node_id, clear_cache=False)
+        document = self._build_document(node)
+        self._documents[node_id] = document
+
+        for term in document.exact_terms:
+            self._exact_index[term].add(node_id)
+
+        for token in document.tokens:
+            self._token_index[token].add(node_id)
+            for length in range(self.prefix_min_length, min(len(token), self.prefix_max_length) + 1):
+                self._prefix_index[token[:length]].add(node_id)
+
+        bisect.insort(self._ordered_node_ids, node_id)
+
+        if clear_cache:
+            self.clear_cache()
+
+    def search(
+        self,
+        query: str,
+        *,
+        limit: int = 20,
+        filters: Optional[Dict[str, Any]] = None,
+    ) -> tuple[List[Tuple[str, float]], Dict[str, Any]]:
+        normalized_query = _normalize_text(query)
+        filters = filters or {}
+        diagnostics: Dict[str, Any] = {
+            "cache_hit": False,
+            "path": "empty",
+            "candidates": 0,
+        }
+        if not normalized_query:
+            return [], diagnostics
+
+        cache_key = self._cache_key(normalized_query, limit, filters)
+        cached = self._cache.get(cache_key)
+        if cached is not None:
+            self._cache.move_to_end(cache_key)
+            diagnostics.update({"cache_hit": True, "path": "cache", "candidates": len(cached)})
+            return list(cached), diagnostics
+
+        query_tokens = _tokenize(normalized_query)
+        exact_ids = set(self._exact_index.get(normalized_query, set()))
+        token_sets: List[set[str]] = []
+        prefix_sets: List[set[str]] = []
+        for token in query_tokens:
+            exact_token_ids = set(self._token_index.get(token, set()))
+            prefix_ids = set(self._prefix_index.get(token, set())) if len(token) >= self.prefix_min_length else set()
+            if exact_token_ids:
+                token_sets.append(exact_token_ids)
+            if prefix_ids:
+                prefix_sets.append(prefix_ids)
+
+        candidate_ids: set[str] = set(exact_ids)
+        if token_sets:
+            intersected = set.intersection(*token_sets)
+            candidate_ids.update(intersected if intersected else set().union(*token_sets))
+        if prefix_sets:
+            candidate_ids.update(set().union(*prefix_sets))
+
+        diagnostics["path"] = "index"
+
+        if not candidate_ids:
+            diagnostics["path"] = "secondary_scan"
+            candidate_ids = self._secondary_scan(normalized_query, limit)
+
+        diagnostics["candidates"] = len(candidate_ids)
+
+        scored: List[Tuple[float, int, int, str]] = []
+        for node_id in candidate_ids:
+            document = self._documents.get(node_id)
+            if document is None or not self._passes_filters(document, filters):
+                continue
+            score = self._score_document(document, normalized_query, query_tokens)
+            if score <= 0:
+                continue
+            token_hits = sum(1 for token in query_tokens if token in document.tokens)
+            exactness = 1 if normalized_query == document.normalized_id or normalized_query in document.exact_terms else 0
+            scored.append((score, exactness, token_hits, node_id))
+
+        top_matches = heapq.nlargest(limit, scored, key=lambda item: (item[0], item[1], item[2], item[3]))
+        results = [(node_id, round(score, 4)) for score, _, _, node_id in top_matches]
+        self._store_cache(cache_key, results)
+        return results, diagnostics
+
+    def _secondary_scan(self, normalized_query: str, limit: int) -> set[str]:
+        matches: set[str] = set()
+        max_hits = max(limit * 20, 200)
+        scanned = 0
+        for node_id in self._ordered_node_ids:
+            if scanned >= self.secondary_scan_limit or len(matches) >= max_hits:
+                break
+            scanned += 1
+            document = self._documents.get(node_id)
+            if document is None:
+                continue
+            if normalized_query in document.primary_text or normalized_query in document.secondary_text:
+                matches.add(node_id)
+        return matches
+
+    def _score_document(
+        self,
+        document: IndexedNodeDocument,
+        normalized_query: str,
+        query_tokens: Tuple[str, ...],
+    ) -> float:
+        score = 0.0
+        if normalized_query == document.normalized_id:
+            score = max(score, 140.0)
+        elif normalized_query in document.exact_terms:
+            score = max(score, 120.0)
+
+        if normalized_query and normalized_query in document.primary_text:
+            score = max(score, 78.0 + min(len(normalized_query), 24) / 10.0)
+        elif normalized_query and normalized_query in document.secondary_text:
+            score = max(score, 26.0 + min(len(normalized_query), 24) / 20.0)
+
+        token_hits = 0
+        prefix_hits = 0
+        for token in query_tokens:
+            if token in document.tokens:
+                token_hits += 1
+            elif len(token) >= self.prefix_min_length and any(candidate.startswith(token) for candidate in document.tokens):
+                prefix_hits += 1
+
+        score += token_hits * 18.0
+        score += prefix_hits * 10.0
+
+        if len(query_tokens) > 1 and token_hits:
+            score += token_hits * 4.0
+
+        return score
+
+    def _passes_filters(self, document: IndexedNodeDocument, filters: Dict[str, Any]) -> bool:
+        filter_type = filters.get("type") or filters.get("node_type")
+        if filter_type and document.node_type != str(filter_type):
+            return False
+
+        min_confidence = _coerce_float(filters.get("min_confidence"))
+        if min_confidence is not None:
+            if document.confidence is None or document.confidence < min_confidence:
+                return False
+
+        tags_filter = filters.get("tags")
+        if tags_filter:
+            if isinstance(tags_filter, str):
+                required_tags = {_normalize_text(tags_filter)}
+            else:
+                required_tags = {
+                    normalized
+                    for normalized in (_normalize_text(tag) for tag in tags_filter)
+                    if normalized
+                }
+            if required_tags and not required_tags.issubset(set(document.tags)):
+                return False
+
+        return True
+
+    def _cache_key(
+        self,
+        normalized_query: str,
+        limit: int,
+        filters: Dict[str, Any],
+    ) -> Tuple[Any, ...]:
+        serialized_filters: List[Tuple[str, Any]] = []
+        for key in sorted(filters.keys()):
+            value = filters[key]
+            if isinstance(value, (list, tuple, set)):
+                serialized_filters.append((key, tuple(sorted(str(item) for item in value))))
+            else:
+                serialized_filters.append((key, str(value)))
+        return normalized_query, limit, tuple(serialized_filters)
+
+    def _store_cache(self, cache_key: Tuple[Any, ...], results: List[Tuple[str, float]]) -> None:
+        self._cache[cache_key] = list(results)
+        self._cache.move_to_end(cache_key)
+        while len(self._cache) > self.cache_size:
+            self._cache.popitem(last=False)
+
+    def _build_document(self, node: Dict[str, Any]) -> IndexedNodeDocument:
+        node_id = str(node.get("id", "")).strip()
+        node_type = str(node.get("type", "entity"))
+        properties = dict(node.get("properties", {}) or {})
+
+        primary_terms: List[str] = []
+        for candidate in (node_id, node.get("content", "")):
+            normalized = _normalize_text(candidate)
+            if normalized:
+                primary_terms.append(normalized)
+
+        for alias_key in _CURATED_ALIAS_KEYS:
+            _collect_text_fragments(properties.get(alias_key), primary_terms, limit=32)
+
+        deduped_primary_terms = tuple(dict.fromkeys(term for term in primary_terms if term))
+        primary_text = " ".join(deduped_primary_terms)
+        tokens = frozenset(_tokenize(primary_text))
+
+        secondary_fragments: List[str] = []
+        for key, value in properties.items():
+            if key in _CURATED_ALIAS_KEYS or key in {"content", "valid_from", "valid_until"}:
+                continue
+            _collect_text_fragments(value, secondary_fragments, limit=48)
+            if len(secondary_fragments) >= 48:
+                break
+
+        secondary_text = " ".join(dict.fromkeys(fragment for fragment in secondary_fragments if fragment))
+        confidence = _coerce_float(properties.get("confidence"))
+
+        raw_tags = properties.get("tags") or []
+        if isinstance(raw_tags, str):
+            raw_tags = [raw_tags]
+        tags = tuple(
+            dict.fromkeys(
+                normalized for normalized in (_normalize_text(tag) for tag in raw_tags) if normalized
+            )
+        )
+
+        return IndexedNodeDocument(
+            node_id=node_id,
+            normalized_id=_normalize_text(node_id),
+            node_type=node_type,
+            exact_terms=frozenset(deduped_primary_terms),
+            tokens=tokens,
+            primary_text=primary_text,
+            secondary_text=secondary_text,
+            confidence=confidence,
+            tags=tags,
+        )

semantica/explorer/session.py

@@ -4,12 +4,15 @@ Semantica Explorer session helpers.
 
 import base64
 import json
+import logging
 import threading
+import time
 import uuid
-from datetime import datetime, UTC
+from datetime import UTC, datetime
 from typing import Any, Dict, Iterable, List, Optional
 
 from ..context.context_graph import ContextGraph, _resolve_edge_identity
+from .search_index import GraphSearchIndex
 
 _KG_AVAILABLE = False
 try:
@@ -28,6 +31,8 @@ try:
 except ImportError:
     pass
 
+logger = logging.getLogger(__name__)
+
 
 class GraphSession:
     """Thread-safe session wrapper around a loaded ``ContextGraph``."""
@@ -35,6 +40,7 @@ class GraphSession:
     def __init__(self, graph: ContextGraph) -> None:
         self.graph = graph
         self._lock = threading.RLock()
+        self._search_index = GraphSearchIndex()
 
         self.annotations: Dict[str, Dict[str, Any]] = {}
 
@@ -46,6 +52,7 @@ class GraphSession:
         self._similarity: Any = None
         self._link_predictor: Any = None
         self._validator: Any = None
+        self.rebuild_search_index()
 
     @classmethod
     def from_file(cls, path: str) -> "GraphSession":
@@ -390,6 +397,28 @@ class GraphSession:
         with self._lock:
             return self.graph.get_neighbors(node_id, hops=depth)
 
+    def rebuild_search_index(self) -> None:
+        with self._lock:
+            normalized_nodes = [
+                self.normalize_node(node.to_dict())
+                for node in self.graph.nodes.values()
+                if node is not None
+            ]
+        self._search_index.rebuild(normalized_nodes)
+
+    def handle_graph_mutation(self, event_type: str, entity_id: str, payload: Dict[str, Any]) -> None:
+        normalized_event = str(event_type or "").upper()
+        if normalized_event in {"ADD_NODE", "UPDATE_NODE"}:
+            normalized_node = self.normalize_node(payload or {})
+            if normalized_node.get("id"):
+                with self._lock:
+                    self._search_index.upsert(normalized_node)
+        elif normalized_event in {"REMOVE_NODE", "DELETE_NODE"}:
+            with self._lock:
+                self._search_index.remove(str(entity_id))
+        elif normalized_event in {"RELOAD_GRAPH", "RESET_GRAPH"}:
+            self.rebuild_search_index()
+
     def search(
         self,
         query: str,
@@ -397,64 +426,34 @@ class GraphSession:
         filters: Optional[Dict[str, Any]] = None,
     ) -> List[Dict[str, Any]]:
         filters = filters or {}
-        try:
-            with self._lock:
-                raw = self.graph.query(query)[:limit]
-        except Exception:
-            raw = []
+        started_at = time.perf_counter()
+        matches, diagnostics = self._search_index.search(query, limit=limit, filters=filters)
 
-        if not raw:
-            nodes, _ = self.get_nodes(search=query, skip=0, limit=max(limit * 5, limit))
-            scored = []
-            lowered_query = query.lower().strip()
-            for node in nodes:
-                haystacks = [
-                    str(node.get("id", "")),
-                    str(node.get("content", "")),
-                    json.dumps(node.get("properties", {}), default=str),
-                ]
-                best_score = 0.0
-                for haystack in haystacks:
-                    lowered = haystack.lower()
-                    if lowered == lowered_query:
-                        best_score = max(best_score, 1.0)
-                    elif lowered_query in lowered:
-                        best_score = max(best_score, min(0.9, len(lowered_query) / max(len(lowered), 1)))
-                if best_score > 0:
-                    scored.append({"node": node, "score": round(best_score, 4)})
-            raw = sorted(scored, key=lambda item: item["score"], reverse=True)[:limit]
-
-        normalized = []
-        for result in raw:
-            result_node = result.get("node", {})
-            node = (
-                self.normalize_node(result_node)
-                if "properties" in result_node or "metadata" in result_node or "content" in result_node
-                else result_node
-            )
-
-            filter_type = filters.get("type") or filters.get("node_type")
-            if filter_type and node["type"] != filter_type:
-                continue
-
-            min_confidence = self._coerce_float(filters.get("min_confidence"))
-            node_confidence = self._coerce_float(node["properties"].get("confidence"))
-            if min_confidence is not None and (
-                node_confidence is None or node_confidence < min_confidence
-            ):
-                continue
-
-            tags_filter = filters.get("tags")
-            if tags_filter:
-                node_tags = node["properties"].get("tags") or []
-                if isinstance(node_tags, str):
-                    node_tags = [node_tags]
-                if not set(tags_filter).issubset(set(node_tags)):
+        normalized_results: List[Dict[str, Any]] = []
+        with self._lock:
+            for node_id, score in matches:
+                raw_node = self.graph.find_node(node_id)
+                if raw_node is None:
                     continue
+                node_payload = raw_node.to_dict() if hasattr(raw_node, "to_dict") else raw_node
+                normalized_results.append(
+                    {
+                        "node": self.normalize_node(node_payload),
+                        "score": score,
+                    }
+                )
 
-            normalized.append({"node": node, "score": result.get("score", 0.0)})
-
-        return normalized[:limit]
+        duration_ms = round((time.perf_counter() - started_at) * 1000, 2)
+        logger.debug(
+            "Explorer search query=%r limit=%s cache_hit=%s path=%s candidates=%s duration_ms=%s",
+            query,
+            limit,
+            diagnostics.get("cache_hit"),
+            diagnostics.get("path"),
+            diagnostics.get("candidates"),
+            duration_ms,
+        )
+        return normalized_results[:limit]
 
     def get_stats(self) -> Dict[str, Any]:
         with self._lock:
@@ -594,8 +593,51 @@ class GraphSession:
 
     def add_nodes(self, nodes: List[Dict[str, Any]]) -> int:
         with self._lock:
-            return self.graph.add_nodes(nodes)
+            added = self.graph.add_nodes(nodes)
+            has_mutation_callback = callable(getattr(self.graph, "mutation_callback", None))
+        if added and not has_mutation_callback:
+            self.rebuild_search_index()
+        return added
 
     def add_edges(self, edges: List[Dict[str, Any]]) -> int:
         with self._lock:
-            return self.graph.add_edges(edges)
+            added = self.graph.add_edges(edges)
+            has_mutation_callback = callable(getattr(self.graph, "mutation_callback", None))
+        if added and not has_mutation_callback:
+            self.rebuild_search_index()
+        return added
+
+    def add_node(
+        self,
+        node_id: str,
+        node_type: str,
+        content: Optional[str] = None,
+        **properties: Any,
+    ) -> bool:
+        with self._lock:
+            added = self.graph.add_node(node_id, node_type, content=content, **properties)
+            has_mutation_callback = callable(getattr(self.graph, "mutation_callback", None))
+        if added and not has_mutation_callback:
+            normalized = self.get_node(node_id)
+            if normalized is not None:
+                self._search_index.upsert(normalized)
+        return added
+
+    def add_edge(
+        self,
+        source_id: str,
+        target_id: str,
+        edge_type: str = "related_to",
+        weight: float = 1.0,
+        **properties: Any,
+    ) -> bool:
+        with self._lock:
+            added = self.graph.add_edge(
+                source_id,
+                target_id,
+                edge_type=edge_type,
+                weight=weight,
+                **properties,
+            )
+            has_mutation_callback = callable(getattr(self.graph, "mutation_callback", None))
+        return added

semantica/export/owl_exporter.py

@@ -328,6 +328,18 @@ class OWLExporter:
         lines.append("</rdf:RDF>")
         return "\n".join(lines)
 
+    @staticmethod
+    def _escape_ttl_str(value: str) -> str:
+        """Escape a string value for safe embedding in a Turtle string literal."""
+        return value.replace("\\", "\\\\").replace('"', '\\"').replace("\n", "\\n").replace("\r", "\\r").replace("\t", "\\t")
+
+    def _ttl_block(self, subject_uri: str, rdf_type: str, predicates: List[str]) -> str:
+        """Build a valid Turtle subject block from accumulated predicate strings."""
+        stmt = f"<{subject_uri}> a {rdf_type}"
+        for pred in predicates:
+            stmt += f" ;\n    {pred}"
+        return stmt + " ."
+
     def _export_owl_turtle(self, ontology: Dict[str, Any], **options) -> str:
         """
         Export ontology to OWL Turtle format.
@@ -342,6 +354,7 @@ class OWLExporter:
         Returns:
             String containing OWL Turtle serialization
         """
+        esc = self._escape_ttl_str
         ontology_uri = ontology.get("uri") or self.ontology_uri
         ontology_name = ontology.get("name", "SemanticaOntology")
         version = ontology.get("version") or self.version
@@ -357,63 +370,73 @@ class OWLExporter:
         lines.append("")
 
         # Ontology declaration
-        lines.append(f"<{ontology_uri}> a owl:Ontology ;")
-        lines.append(f'    rdfs:label "{ontology_name}" ;')
-        lines.append(f'    owl:versionInfo "{version}" .')
-        if ontology.get("description"):
-            lines.append(f'    rdfs:comment "{ontology.get("description")}" ;')
+        onto_predicates = [
+            f'rdfs:label "{esc(ontology_name)}"',
+            f'owl:versionInfo "{esc(version)}"',
+        ]
+        description = ontology.get("description")
+        if description:
+            onto_predicates.append(f'rdfs:comment "{esc(description)}"')
+        lines.append(self._ttl_block(ontology_uri, "owl:Ontology", onto_predicates))
         lines.append("")
 
         # Classes
-        classes = ontology.get("classes", [])
-        for cls in classes:
+        for cls in ontology.get("classes", []):
             class_uri = cls.get("uri") or cls.get("id", "")
             class_name = cls.get("name") or cls.get("label", "")
-
-            lines.append(f"<{class_uri}> a owl:Class ;")
-            lines.append(f'    rdfs:label "{class_name}" .')
-
-            if cls.get("comment"):
-                lines.append(f'    rdfs:comment "{cls.get("comment")}" ;')
-
-            if cls.get("subClassOf"):
-                parent = cls.get("subClassOf")
-                lines.append(f"    rdfs:subClassOf <{parent}> ;")
-
-            # Remove trailing semicolon and add period
-            if lines[-1].endswith(" ;"):
-                lines[-1] = lines[-1].rstrip(" ;") + " ."
-            else:
-                lines.append(" .")
+            predicates = [f'rdfs:label "{esc(class_name)}"']
+            comment = cls.get("comment")
+            if comment:
+                predicates.append(f'rdfs:comment "{esc(comment)}"')
+            sub_class = cls.get("subClassOf")
+            if sub_class:
+                predicates.append(f"rdfs:subClassOf <{sub_class}>")
+            equiv = cls.get("equivalentClass")
+            if equiv:
+                predicates.append(f"owl:equivalentClass <{equiv}>")
+            lines.append(self._ttl_block(class_uri, "owl:Class", predicates))
             lines.append("")
 
         # Object properties
-        object_properties = ontology.get("object_properties", [])
-        for prop in object_properties:
+        for prop in ontology.get("object_properties", []):
             prop_uri = prop.get("uri") or prop.get("id", "")
             prop_name = prop.get("name") or prop.get("label", "")
-
-            lines.append(f"<{prop_uri}> a owl:ObjectProperty ;")
-            lines.append(f'    rdfs:label "{prop_name}" .')
-
-            if prop.get("domain"):
-                domain = prop.get("domain")
+            predicates = [f'rdfs:label "{esc(prop_name)}"']
+            comment = prop.get("comment")
+            if comment:
+                predicates.append(f'rdfs:comment "{esc(comment)}"')
+            domain = prop.get("domain")
+            if domain:
                 if isinstance(domain, list):
                     for d in domain:
-                        lines.append(f"    rdfs:domain <{d}> ;")
+                        predicates.append(f"rdfs:domain <{d}>")
                 else:
-                    lines.append(f"    rdfs:domain <{domain}> ;")
-
-            if prop.get("range"):
-                range_val = prop.get("range")
+                    predicates.append(f"rdfs:domain <{domain}>")
+            range_val = prop.get("range")
+            if range_val:
                 if isinstance(range_val, list):
                     for r in range_val:
-                        lines.append(f"    rdfs:range <{r}> ;")
+                        predicates.append(f"rdfs:range <{r}>")
                 else:
-                    lines.append(f"    rdfs:range <{range_val}> ;")
+                    predicates.append(f"rdfs:range <{range_val}>")
+            lines.append(self._ttl_block(prop_uri, "owl:ObjectProperty", predicates))
+            lines.append("")
 
-            if lines[-1].endswith(" ;"):
-                lines[-1] = lines[-1].rstrip(" ;") + " ."
+        # Data properties
+        for prop in ontology.get("data_properties", []):
+            prop_uri = prop.get("uri") or prop.get("id", "")
+            prop_name = prop.get("name") or prop.get("label", "")
+            predicates = [f'rdfs:label "{esc(prop_name)}"']
+            comment = prop.get("comment")
+            if comment:
+                predicates.append(f'rdfs:comment "{esc(comment)}"')
+            domain = prop.get("domain")
+            if domain:
+                predicates.append(f"rdfs:domain <{domain}>")
+            range_type = prop.get("range")
+            if range_type:
+                predicates.append(f"rdfs:range xsd:{range_type}")
+            lines.append(self._ttl_block(prop_uri, "owl:DatatypeProperty", predicates))
             lines.append("")
 
         return "\n".join(lines)

semantica/kg/__init__.py

@@ -126,12 +126,14 @@ from .temporal_query import (
     TemporalPatternDetector,
     TemporalVersionManager,
 )
+from .knowledge_graph import KnowledgeGraph
 from .temporal_model import BiTemporalFact, TemporalBound
 from .temporal_normalizer import TemporalNormalizer
 from .temporal_query_rewriter import TemporalQueryRewriter, TemporalQueryResult
 
 __all__ = [
     # Core Classes
+    "KnowledgeGraph",
     "GraphBuilder",
     "GraphBuilderWithProvenance",
     "EntityResolver",

semantica/kg/knowledge_graph.py

@@ -0,0 +1,46 @@
+"""
+KnowledgeGraph dataclass — canonical in-memory representation.
+
+This is the formal type produced by the Semantica KG pipeline and consumed
+by visualizers, exporters, and other downstream components.  It is a thin,
+immutable-friendly wrapper around three plain collections so that isinstance
+checks, type hints, and IDEs can surface the type rather than relying on
+bare dicts.
+
+Keeping this in its own file avoids circular imports between the kg and
+visualization sub-packages.
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from typing import Any, Dict, List
+
+
+@dataclass
+class KnowledgeGraph:
+    """
+    Canonical in-memory knowledge graph.
+
+    Attributes:
+        entities:      List of entity dicts with at minimum ``id`` and ``type`` keys.
+        relationships: List of relationship dicts with at minimum ``source``,
+                       ``target``, and ``type`` keys.
+        metadata:      Arbitrary graph-level metadata (e.g. build timestamps,
+                       entity-resolution flags).
+    """
+
+    entities: List[Dict[str, Any]] = field(default_factory=list)
+    relationships: List[Dict[str, Any]] = field(default_factory=list)
+    metadata: Dict[str, Any] = field(default_factory=dict)
+
+    # ------------------------------------------------------------------
+    # Convenience helpers
+    # ------------------------------------------------------------------
+
+    def __len__(self) -> int:
+        """Return the number of entities (mirrors the most common 'size' query)."""
+        return len(self.entities)
+
+    def __bool__(self) -> bool:
+        return bool(self.entities or self.relationships)

semantica/kg/path_finder.py

@@ -104,7 +104,8 @@ class PathFinder:
         source: str,
         target: str,
         weight_attribute: str = "weight",
-        default_weight: float = 1.0
+        default_weight: float = 1.0,
+        directed: bool = True
     ) -> List[str]:
         """
         Find shortest path using Dijkstra's algorithm.
@@ -125,32 +126,34 @@ class PathFinder:
         """
         try:
             self.logger.info(f"Finding Dijkstra shortest path from {source} to {target}")
-            
+
             # Validate nodes exist
             if not self._node_exists(graph, source):
                 raise ValueError(f"Source node {source} not found")
             if not self._node_exists(graph, target):
                 raise ValueError(f"Target node {target} not found")
-            
+
+            traversal_graph = graph if directed else self._make_undirected_view(graph)
+
             # Dijkstra's algorithm
             distances = {source: 0.0}
             previous = {}
             priority_queue = [(0.0, source)]
             visited = set()
-            
+
             while priority_queue:
                 current_distance, current_node = heapq.heappop(priority_queue)
-                
+
                 if current_node in visited:
                     continue
-                
+
                 visited.add(current_node)
-                
+
                 if current_node == target:
                     break
-                
+
                 # Explore neighbors
-                for neighbor, edge_data in self._get_neighbors(graph, current_node):
+                for neighbor, edge_data in self._get_neighbors(traversal_graph, current_node):
                     if neighbor in visited:
                         continue
                     
@@ -350,44 +353,48 @@ class PathFinder:
         self,
         graph: Any,
         source: str,
-        target: str
+        target: str,
+        directed: bool = True
     ) -> List[str]:
         """
         Find shortest path using BFS (unweighted).
-        
+
         Args:
             graph: Graph object (NetworkX or similar)
             source: Source node ID
             target: Target node ID
-            
+            directed: If False, treat the graph as undirected for traversal
+
         Returns:
             List of node IDs representing the shortest path
-            
+
         Raises:
             ValueError: If source or target not found
         """
         try:
             self.logger.info(f"Finding BFS shortest path from {source} to {target}")
-            
+
             # Validate nodes exist
             if not self._node_exists(graph, source):
                 raise ValueError(f"Source node {source} not found")
             if not self._node_exists(graph, target):
                 raise ValueError(f"Target node {target} not found")
-            
+
+            traversal_graph = graph if directed else self._make_undirected_view(graph)
+
             # BFS algorithm
             queue = deque([(source, [source])])
             visited = {source}
-            
+
             while queue:
                 current, path = queue.popleft()
-                
+
                 if current == target:
                     self.logger.info(f"Found BFS path of length {len(path)}")
                     return path
-                
+
                 # Explore neighbors
-                for neighbor, _ in self._get_neighbors(graph, current):
+                for neighbor, _ in self._get_neighbors(traversal_graph, current):
                     if neighbor not in visited:
                         visited.add(neighbor)
                         queue.append((neighbor, path + [neighbor]))
@@ -564,6 +571,18 @@ class PathFinder:
                 return False
         return False
     
+    def _make_undirected_view(self, graph: Any) -> Any:
+        """Return an undirected view of the graph for bidirectional traversal.
+
+        For NetworkX directed graphs this calls ``to_undirected()``, which
+        preserves all edge attributes.  For graph types that have no such
+        method the original object is returned as a fallback — callers that
+        already expose undirected neighbors will still work correctly.
+        """
+        if hasattr(graph, "to_undirected"):
+            return graph.to_undirected()
+        return graph
+
     def _get_neighbors(self, graph: Any, node: str) -> List[Tuple[str, Any]]:
         """Get neighbors of a node with edge data."""
         neighbors = []

semantica/semantic_extract/providers.py

@@ -397,6 +397,7 @@ class BaseProvider:
                         create_kwargs["response_format"] = {"type": "json_object"}
                     
                     response = client.chat.completions.create(**create_kwargs)
+                    verbose_mode = kwargs.get("verbose", False) or self.config.get("verbose", False)
                     if verbose_mode:
                         import sys
                         print(f"    [BaseProvider.generate_typed] Typed response received via instructor ({provider_name}).", flush=True, file=sys.stdout)
@@ -939,20 +940,22 @@ class DeepSeekProvider(BaseProvider):
     def __init__(self, api_key: Optional[str] = None, model: str = "deepseek-chat", **kwargs):
         super().__init__(**kwargs)
         self.api_key = api_key or config.get_api_key("deepseek")
+        self.base_url = "https://api.deepseek.com/v1"
         self.model = model
+        self.base_url = "https://api.deepseek.com/v1"
         self.client = None
         self._init_client()
 
     def _init_client(self):
         try:
-            import deepseek  # type: ignore[import-untyped]
+            from openai import OpenAI
 
             if self.api_key:
-                self.client = deepseek.Client(api_key=self.api_key)
+                self.client = OpenAI(api_key=self.api_key, base_url=self.base_url)
         except (ImportError, OSError):
             self.client = None
             self.logger.warning(
-                "deepseek library not installed. Install with: pip install semantica[llm-deepseek]"
+                "openai library not installed. Install with: pip install semantica[llm-openai]"
             )
 
     def is_available(self) -> bool:

semantica/server.py

@@ -188,10 +188,22 @@ async def serve_spa(full_path: str):
     if full_path.startswith("api/"):
         raise HTTPException(status_code=404, detail="API route not found")
 
+    # Root path — serve index.html if built, otherwise a welcome JSON response
+    if full_path in ("", "/"):
+        index_file = STATIC_DIR / "index.html"
+        if index_file.is_file():
+            return FileResponse(index_file)
+        return JSONResponse({
+            "name": "Semantica Knowledge Explorer",
+            "version": __version__,
+            "message": "Welcome to Semantica. The frontend is not built yet — run `npm run build` inside the explorer/ directory, or open the Vite dev server at http://localhost:5173.",
+            "docs": "/docs",
+            "health": "/health",
+        })
+
     normalized_path = os.path.normpath(full_path)
     if (
-        normalized_path in ("", ".")
-        or os.path.isabs(normalized_path)
+        os.path.isabs(normalized_path)
         or normalized_path == ".."
         or normalized_path.startswith(".." + os.sep)
     ):
@@ -200,7 +212,7 @@ async def serve_spa(full_path: str):
     # Ensure join remains relative to STATIC_DIR even if input includes leading separators
     safe_rel_path = normalized_path.lstrip("/\\")
     rel_parts = Path(safe_rel_path).parts
-    if any(part in ("", ".", "..") for part in rel_parts):
+    if any(part in (".", "..") for part in rel_parts):
         raise HTTPException(status_code=400, detail="Invalid path")
 
     static_dir_resolved = STATIC_DIR.resolve()

semantica/utils/helpers.py

@@ -562,3 +562,25 @@ def retry_on_error(
         return wrapper
 
     return decorator
+
+
+def classify_path_distance(hop_count: int) -> str:
+    """Classify a path hop count into a human-readable distance band.
+
+    Bands:
+        "direct"    — 0–1 hops (single edge or self)
+        "near"      — 2–3 hops (closely related)
+        "mid-range" — 4–6 hops (reachable but separated)
+        "distant"   — 7+ hops (weakly coupled)
+
+    This is the single source of truth for distance-band thresholds used by
+    both the Explorer API (PathResponse.distance_band) and the KGVisualizer
+    (highlight_path edge styling).
+    """
+    if hop_count <= 1:
+        return "direct"
+    if hop_count <= 3:
+        return "near"
+    if hop_count <= 6:
+        return "mid-range"
+    return "distant"

semantica/visualization/kg_visualizer.py

@@ -53,6 +53,15 @@ except (ImportError, OSError):
 
 from ..utils.exceptions import ProcessingError
 from ..utils.logging import get_logger
+
+# Optional import — keeps the visualizer usable even if the kg sub-package
+# is not installed, and avoids circular-import risk at module level.
+try:
+    from ..kg.knowledge_graph import KnowledgeGraph as _KnowledgeGraph
+except Exception:  # pragma: no cover
+    _KnowledgeGraph = None  # type: ignore[assignment,misc]
+
+from ..utils.helpers import classify_path_distance
 from ..utils.progress_tracker import get_progress_tracker
 from .utils.color_schemes import ColorPalette, ColorScheme
 from .utils.export_formats import (
@@ -118,18 +127,45 @@ class KGVisualizer:
                 "Install with: pip install plotly"
             )
 
-    def _normalize_graph(self, graph: Any) -> Dict[str, Any]:
+    def _convert_knowledge_graph(self, kg: Any) -> Dict[str, Any]:
         """
-        Normalize graph input to the expected dict format.
+        Convert a KnowledgeGraph instance to the internal dict format.
 
-        Accepts either:
-        - A dict with "entities" and "relationships" keys (canonical format)
-        - Any object that exposes .entities and .relationships attributes
-          (e.g. a KnowledgeGraph dataclass returned by GraphBuilder.build())
+        Non-mutating. Preserves node types, labels, properties, edge types,
+        weights, and direction.
+
+        Args:
+            kg: A ``KnowledgeGraph`` instance.
 
         Returns:
             Dict with "entities", "relationships", and "metadata" keys.
         """
+        entities = getattr(kg, "entities", None) or []
+        relationships = getattr(kg, "relationships", None) or []
+        metadata = getattr(kg, "metadata", None) or {}
+        return {
+            "entities": list(entities),
+            "relationships": list(relationships),
+            "metadata": dict(metadata),
+        }
+
+    def _normalize_graph(self, graph: Any) -> Dict[str, Any]:
+        """
+        Normalize graph input to the expected dict format.
+
+        Accepts:
+        - A ``KnowledgeGraph`` instance (routed through ``_convert_knowledge_graph``)
+        - A dict with "entities" and "relationships" keys (canonical format)
+        - Any object that exposes .entities and .relationships attributes
+          (duck-typed, e.g. custom dataclasses)
+
+        Returns:
+            Dict with "entities", "relationships", and "metadata" keys.
+        """
+        # Explicit fast-path for the formal KnowledgeGraph type
+        if _KnowledgeGraph is not None and isinstance(graph, _KnowledgeGraph):
+            return self._convert_knowledge_graph(graph)
+
         if isinstance(graph, dict):
             return graph
 
@@ -156,6 +192,7 @@ class KGVisualizer:
         node_color_by: str = "type",
         node_size_by: Optional[str] = None,
         hover_data: Optional[List[str]] = None,
+        highlight_path: Optional[List[str]] = None,
         **options,
     ) -> Optional[Any]:
         """
@@ -177,6 +214,9 @@ class KGVisualizer:
             node_color_by: Property to map to node color (default: "type")
             node_size_by: Property to map to node size (default: fixed)
             hover_data: List of properties to show in hover tooltip
+            highlight_path: Optional ordered list of node IDs forming a path to
+                highlight with distance-aware edge styling (opacity and stroke
+                weight reflect hop count along the path).
             **options: Additional visualization options
 
         Returns:
@@ -226,13 +266,14 @@ class KGVisualizer:
                 tracking_id, message="Generating visualization..."
             )
             result = self._visualize_network_plotly(
-                nodes, 
-                edges, 
-                output, 
-                file_path, 
+                nodes,
+                edges,
+                output,
+                file_path,
                 node_color_by=node_color_by,
                 node_size_by=node_size_by,
                 hover_data=hover_data,
+                highlight_path=highlight_path,
                 **options
             )
 
@@ -529,6 +570,21 @@ class KGVisualizer:
 
         return edges
 
+    @staticmethod
+    def _path_edge_style(distance_band: str) -> Tuple[float, float]:
+        """Return (opacity, width) for a path edge based on its distance band.
+
+        Bands come from ``classify_path_distance`` in ``utils.helpers`` — the
+        single source of truth for hop-count thresholds.
+        """
+        if distance_band == "direct":
+            return (1.0, 4.0)
+        if distance_band == "near":
+            return (0.85, 3.0)
+        if distance_band == "mid-range":
+            return (0.6, 2.0)
+        return (0.35, 1.5)  # "distant"
+
     def _visualize_network_plotly(
         self,
         nodes: List[Dict[str, Any]],
@@ -538,6 +594,7 @@ class KGVisualizer:
         node_color_by: str = "type",
         node_size_by: Optional[str] = None,
         hover_data: Optional[List[str]] = None,
+        highlight_path: Optional[List[str]] = None,
         **options,
     ) -> Optional[Any]:
         """Create Plotly network visualization."""
@@ -640,47 +697,73 @@ class KGVisualizer:
                 
             node_text.append(text)
 
-        # Prepare edge traces
-        edge_x = []
-        edge_y = []
-        
+        # Build path edge lookup for highlight_path support
+        path_edge_set: set = set()
+        path_distance_band = "direct"
+        if highlight_path and len(highlight_path) >= 2:
+            path_hop_count = len(highlight_path) - 1
+            path_distance_band = classify_path_distance(path_hop_count)
+            # Only add the directed edges that actually form the path (A→B, not B→A).
+            # Adding the reverse would incorrectly highlight unrelated back-edges.
+            for i in range(path_hop_count):
+                path_edge_set.add((highlight_path[i], highlight_path[i + 1]))
+            # Warn if any path node has no layout position (silent highlight failure).
+            missing = [n for n in highlight_path if n not in pos]
+            if missing:
+                self.logger.warning(
+                    "highlight_path contains node IDs not found in the graph: %s",
+                    missing,
+                )
+
+        path_opacity, path_width = self._path_edge_style(path_distance_band)
+
+        # Prepare edge traces — split into background (non-path) and path edges
+        edge_x: List = []
+        edge_y: List = []
+        path_edge_x: List = []
+        path_edge_y: List = []
+
         # Prepare edge label traces and annotations (for arrows)
         edge_label_x = []
         edge_label_y = []
         edge_label_text = []
         annotations = []
-        
+
         # Limit detailed edge rendering for performance if graph is too large
         show_detailed_edges = len(edges) < 500
-        
+
         for edge in edges:
             source_pos = pos.get(edge["source"])
             target_pos = pos.get(edge["target"])
             if source_pos and target_pos:
                 x0, y0 = source_pos
                 x1, y1 = target_pos
-                edge_x.extend([x0, x1, None])
-                edge_y.extend([y0, y1, None])
-                
+
+                is_path_edge = (edge["source"], edge["target"]) in path_edge_set
+                if is_path_edge:
+                    path_edge_x.extend([x0, x1, None])
+                    path_edge_y.extend([y0, y1, None])
+                else:
+                    edge_x.extend([x0, x1, None])
+                    edge_y.extend([y0, y1, None])
+
                 if show_detailed_edges:
                     # Calculate midpoint for label
                     mx, my = (x0 + x1) / 2, (y0 + y1) / 2
-                    
+
                     if edge.get("label"):
                         edge_label_x.append(mx)
                         edge_label_y.append(my)
                         edge_label_text.append(edge["label"])
-                    
+
                     # Add arrow annotation
-                    # Adjust arrow to point slightly before the node to avoid overlap with node marker
-                    # This is approximate; precise calculation requires node size
                     annotations.append(
                         dict(
                             ax=x0, ay=y0, axref='x', ayref='y',
                             x=x1, y=y1, xref='x', yref='y',
                             arrowhead=2, arrowsize=1, arrowwidth=1,
                             arrowcolor="#888", opacity=0.6,
-                            standoff=15 # Distance from target node
+                            standoff=15
                         )
                     )
 
@@ -693,9 +776,22 @@ class KGVisualizer:
             showlegend=False,
             opacity=0.5
         )
-        
+
         traces = [edge_trace]
-        
+
+        # Overlay highlighted path edges with distance-aware styling
+        if path_edge_x:
+            path_trace = go.Scatter(
+                x=path_edge_x,
+                y=path_edge_y,
+                line=dict(width=path_width, color="#e05c00"),
+                hoverinfo="none",
+                mode="lines",
+                showlegend=False,
+                opacity=path_opacity,
+            )
+            traces.append(path_trace)
+
         if show_detailed_edges and edge_label_text:
             edge_label_trace = go.Scatter(
                 x=edge_label_x,

tests/explorer/test_explorer_api.py

@@ -4,6 +4,7 @@ import json
 from pathlib import Path
 import uuid
 
+import networkx as nx
 import pytest
 
 from semantica.context.context_graph import ContextGraph
@@ -35,6 +36,16 @@ def _build_sample_graph() -> ContextGraph:
     graph.add_node("javascript", node_type="language", content="JavaScript programming language", x=100, y=120)
     graph.add_node("web_dev", node_type="concept", content="Web Development", x=24, y=30)
     graph.add_node("ml", node_type="concept", content="Machine Learning", x=45, y=60)
+    graph.add_node(
+        "metformin",
+        node_type="drug",
+        content="Metformin",
+        aliases=["Glucophage"],
+        confidence="0.97",
+        tags=["drug", "featured"],
+        x=22,
+        y=33,
+    )
     graph.add_node(
         "decision_1",
         node_type="decision",
@@ -243,6 +254,74 @@ class TestSearchAndStats:
         assert payload["total"] >= 1
         assert all(item["node"]["type"] == "language" for item in payload["results"])
 
+    def test_search_exact_and_prefix(self, client):
+        exact_response = client.post(
+            "/api/graph/search",
+            json={"query": "Metformin", "limit": 5},
+        )
+        assert exact_response.status_code == 200
+        exact_payload = exact_response.json()
+        assert exact_payload["results"][0]["node"]["id"] == "metformin"
+
+        prefix_response = client.post(
+            "/api/graph/search",
+            json={"query": "metf", "limit": 5},
+        )
+        assert prefix_response.status_code == 200
+        prefix_payload = prefix_response.json()
+        assert any(item["node"]["id"] == "metformin" for item in prefix_payload["results"])
+
+    def test_search_filters_and_cache_stability(self, client):
+        body = {
+            "query": "framework",
+            "filters": {"type": "decision", "min_confidence": 0.8},
+            "limit": 5,
+        }
+        first_response = client.post("/api/graph/search", json=body)
+        second_response = client.post("/api/graph/search", json=body)
+
+        assert first_response.status_code == 200
+        assert second_response.status_code == 200
+        assert first_response.json() == second_response.json()
+        results = first_response.json()["results"]
+        assert [item["node"]["id"] for item in results] == ["decision_1"]
+
+    def test_search_sees_new_nodes_after_mutation(self, client):
+        session = client.app.state.session
+        assert session.add_node(
+            "metformin_hcl",
+            "drug",
+            content="Metformin Hydrochloride",
+            aliases=["Glucophage XR"],
+            confidence="0.93",
+        )
+
+        response = client.post(
+            "/api/graph/search",
+            json={"query": "glucophage", "limit": 10},
+        )
+        assert response.status_code == 200
+        result_ids = [item["node"]["id"] for item in response.json()["results"]]
+        assert "metformin" in result_ids
+        assert "metformin_hcl" in result_ids
+
+    def test_search_secondary_scan_fallback_matches_non_curated_properties(self, client):
+        session = client.app.state.session
+        assert session.add_node(
+            "fallback_node",
+            "entity",
+            content="Alpha",
+            description="rareterm",
+        )
+
+        response = client.post(
+            "/api/graph/search",
+            json={"query": "rareterm", "limit": 10},
+        )
+        assert response.status_code == 200
+        result_ids = [item["node"]["id"] for item in response.json()["results"]]
+        assert "fallback_node" in result_ids
+
     def test_stats(self, client):
         response = client.get("/api/graph/stats")
         assert response.status_code == 200
@@ -372,7 +451,7 @@ class TestEnrichment:
 
     def test_extract(self, client):
         response = client.post("/api/enrich/extract", json={"text": "Alice works at Acme Corp."})
-        assert response.status_code in (200, 422)
+        assert response.status_code in (200, 422, 503)
 
     def test_link_prediction(self, client):
         response = client.post("/api/enrich/links", json={"node_id": "python", "top_n": 5})
@@ -638,3 +717,177 @@ class TestGenericGraphFileLoading:
             assert repeat.status_code == 200
             repeat_ids = [edge["id"] for edge in repeat.json()["edges"]]
             assert repeat_ids == ["edge-alpha", "edge-beta"]
+
+
+# ---------------------------------------------------------------------------
+# Bidirectional path-finding tests (issue #469)
+# ---------------------------------------------------------------------------
+
+def _make_path_session() -> GraphSession:
+    """Return a GraphSession whose build_graph_dict yields an nx.DiGraph with A→B only.
+
+    GraphSession wraps a ContextGraph (required by create_app), but we patch
+    build_graph_dict so PathFinder receives an actual NetworkX DiGraph — the
+    graph type the Explorer is designed to traverse for path queries.
+    """
+    cg = ContextGraph(advanced_analytics=False)
+    cg.add_node("A", node_type="entity", content="Node A")
+    cg.add_node("B", node_type="entity", content="Node B")
+    cg.add_edge("A", "B", edge_type="connects")
+
+    session = GraphSession(cg)
+
+    # Patch build_graph_dict to return the directed NetworkX graph that
+    # PathFinder needs.  The ContextGraph dict format is not traversable by
+    # PathFinder; this mimics how a KG-backed session would expose the graph.
+    digraph = nx.DiGraph()
+    digraph.add_edge("A", "B")
+    session.build_graph_dict = lambda node_ids=None: digraph  # type: ignore[method-assign]
+
+    return session
+
+
+@pytest.fixture
+def path_client():
+    session = _make_path_session()
+    app = create_app(session=session)
+    with TestClient(app) as c:
+        yield c
+
+
+class TestBidirectionalPathRoute:
+    """API-level tests for directed=true/false on GET /api/graph/node/{id}/path."""
+
+    # ------------------------------------------------------------------
+    # directed=true (default) — existing directed-only behaviour
+    # ------------------------------------------------------------------
+
+    def test_directed_true_forward_path_found(self, path_client):
+        """A→B exists: forward query with directed=true must succeed."""
+        resp = path_client.get("/api/graph/node/A/path?target=B&directed=true")
+        assert resp.status_code == 200
+        body = resp.json()
+        assert body["path"] == ["A", "B"]
+        assert body["directed"] is True
+
+    def test_directed_true_reverse_returns_404(self, path_client):
+        """Only A→B exists: reverse query with directed=true must return 404."""
+        resp = path_client.get("/api/graph/node/B/path?target=A&directed=true")
+        assert resp.status_code == 404
+
+    def test_default_param_reverse_returns_404(self, path_client):
+        """Omitting directed= must preserve current directed behaviour (404 for reverse)."""
+        resp = path_client.get("/api/graph/node/B/path?target=A")
+        assert resp.status_code == 404
+
+    # ------------------------------------------------------------------
+    # directed=false — new undirected traversal
+    # ------------------------------------------------------------------
+
+    def test_directed_false_reverse_path_found(self, path_client):
+        """directed=false must find B→A even though only A→B exists."""
+        resp = path_client.get("/api/graph/node/B/path?target=A&directed=false")
+        assert resp.status_code == 200
+        body = resp.json()
+        assert body["path"] == ["B", "A"]
+        assert body["directed"] is False
+
+    def test_directed_false_forward_path_found(self, path_client):
+        """directed=false must not break the natural A→B direction."""
+        resp = path_client.get("/api/graph/node/A/path?target=B&directed=false")
+        assert resp.status_code == 200
+        body = resp.json()
+        assert body["path"] == ["A", "B"]
+        assert body["directed"] is False
+
+    # ------------------------------------------------------------------
+    # Algorithm variants
+    # ------------------------------------------------------------------
+
+    def test_dijkstra_directed_false_reverse(self, path_client):
+        resp = path_client.get(
+            "/api/graph/node/B/path?target=A&algorithm=dijkstra&directed=false"
+        )
+        assert resp.status_code == 200
+        body = resp.json()
+        assert body["path"] == ["B", "A"]
+        assert body["algorithm"] == "dijkstra"
+        assert body["directed"] is False
+
+    def test_dijkstra_directed_true_reverse_returns_404(self, path_client):
+        resp = path_client.get(
+            "/api/graph/node/B/path?target=A&algorithm=dijkstra&directed=true"
+        )
+        assert resp.status_code == 404
+
+    # ------------------------------------------------------------------
+    # PathResponse schema
+    # ------------------------------------------------------------------
+
+    def test_response_schema_includes_directed_field(self, path_client):
+        """PathResponse must always include the directed field."""
+        resp = path_client.get("/api/graph/node/A/path?target=B")
+        assert resp.status_code == 200
+        body = resp.json()
+        assert "directed" in body
+
+    def test_response_directed_reflects_query_param(self, path_client):
+        resp_true = path_client.get("/api/graph/node/A/path?target=B&directed=true")
+        resp_false = path_client.get("/api/graph/node/A/path?target=B&directed=false")
+        assert resp_true.json()["directed"] is True
+        assert resp_false.json()["directed"] is False
+
+    # ------------------------------------------------------------------
+    # hop_count and distance_band — issue #472
+    # ------------------------------------------------------------------
+
+    def test_response_includes_hop_count_and_distance_band(self, path_client):
+        """PathResponse must include hop_count and distance_band fields."""
+        resp = path_client.get("/api/graph/node/A/path?target=B")
+        assert resp.status_code == 200
+        body = resp.json()
+        assert "hop_count" in body
+        assert "distance_band" in body
+
+    def test_one_hop_path_is_direct(self, path_client):
+        """A single-edge path (1 hop) must return distance_band='direct'."""
+        resp = path_client.get("/api/graph/node/A/path?target=B")
+        assert resp.status_code == 200
+        body = resp.json()
+        assert body["hop_count"] == 1
+        assert body["distance_band"] == "direct"
+
+
+# ---------------------------------------------------------------------------
+# _classify_distance unit tests — issue #472
+# ---------------------------------------------------------------------------
+
+from semantica.utils.helpers import classify_path_distance
+
+
+class TestClassifyDistance:
+    """Unit tests covering all four band boundaries."""
+
+    def test_zero_hops_is_direct(self):
+        assert classify_path_distance(0) == "direct"
+
+    def test_one_hop_is_direct(self):
+        assert classify_path_distance(1) == "direct"
+
+    def test_two_hops_is_near(self):
+        assert classify_path_distance(2) == "near"
+
+    def test_three_hops_is_near(self):
+        assert classify_path_distance(3) == "near"
+
+    def test_four_hops_is_mid_range(self):
+        assert classify_path_distance(4) == "mid-range"
+
+    def test_six_hops_is_mid_range(self):
+        assert classify_path_distance(6) == "mid-range"
+
+    def test_seven_hops_is_distant(self):
+        assert classify_path_distance(7) == "distant"
+
+    def test_large_hop_count_is_distant(self):
+        assert classify_path_distance(20) == "distant"

tests/explorer/test_provenance_route.py

@@ -0,0 +1,74 @@
+"""Unit tests for explorer provenance route helpers."""
+
+from types import SimpleNamespace
+
+from semantica.explorer.routes.provenance import _build_provenance, _render_markdown
+
+
+def _make_session_with_chain() -> SimpleNamespace:
+    """Build a minimal session-like object for Source -> Intermediate -> node_id."""
+    nodes = {
+        "Source": SimpleNamespace(node_type="entity", content="Source"),
+        "Intermediate": SimpleNamespace(node_type="entity", content="Intermediate"),
+        "node_id": SimpleNamespace(node_type="entity", content="Target"),
+    }
+    edges = [
+        SimpleNamespace(source_id="Source", target_id="Intermediate", edge_type="related_to"),
+        SimpleNamespace(source_id="Intermediate", target_id="node_id", edge_type="related_to"),
+    ]
+    graph = SimpleNamespace(nodes=nodes, edges=edges)
+    return SimpleNamespace(graph=graph)
+
+
+def test_build_provenance_direction_classification_chain():
+    session = _make_session_with_chain()
+
+    data = _build_provenance(session, "node_id")
+
+    node_ids = {node["id"] for node in data["nodes"]}
+    assert "Source" in node_ids
+    assert "Intermediate" in node_ids
+
+    edge_by_pair = {(edge["source"], edge["target"]): edge for edge in data["edges"]}
+
+    assert edge_by_pair[("Intermediate", "node_id")]["direction"] == "upstream"
+    assert edge_by_pair[("Source", "Intermediate")]["direction"] != "downstream"
+
+
+def test_render_markdown_groups_edges_by_direction():
+    report = {
+        "node_id": "node_id",
+        "label": "Target",
+        "type": "entity",
+        "properties": {},
+        "lineage": {
+            "nodes": [
+                {"id": "Source", "prov_type": "Entity", "label": "Source"},
+                {"id": "Intermediate", "prov_type": "Entity", "label": "Intermediate"},
+                {"id": "node_id", "prov_type": "Entity", "label": "Target"},
+            ],
+            "edges": [
+                {
+                    "id": "Intermediate-node_id",
+                    "source": "Intermediate",
+                    "target": "node_id",
+                    "label": "related_to",
+                    "direction": "upstream",
+                },
+                {
+                    "id": "Source-Intermediate",
+                    "source": "Source",
+                    "target": "Intermediate",
+                    "label": "related_to",
+                    "direction": "lateral",
+                },
+            ],
+        },
+    }
+
+    markdown = _render_markdown(report)
+
+    assert "## Upstream" in markdown
+    assert "## Lateral" in markdown
+    assert "`Intermediate` -[related_to]-> `node_id`" in markdown
+    assert "`Source` -[related_to]-> `Intermediate`" in markdown

tests/export/test_owl_exporter.py

@@ -0,0 +1,549 @@
+"""Tests for OWLExporter._export_owl_turtle fixes (issue #478).
+
+Bug 1: invalid Turtle when subClassOf/domain/range present (predicates appended
+       after a closing period).
+Bug 2: data_properties silently dropped from Turtle output.
+"""
+
+import pytest
+from semantica.export import OWLExporter
+
+
+# ---------------------------------------------------------------------------
+# Shared fixtures
+# ---------------------------------------------------------------------------
+
+@pytest.fixture
+def exporter():
+    return OWLExporter()
+
+
+@pytest.fixture
+def full_ontology():
+    return {
+        "uri": "http://example.org/onto",
+        "name": "TestOntology",
+        "description": "A test ontology",
+        "classes": [
+            {
+                "uri": "http://example.org/Person",
+                "name": "Person",
+            },
+            {
+                "uri": "http://example.org/Employee",
+                "name": "Employee",
+                "comment": "A person who is employed",
+                "subClassOf": "http://example.org/Person",
+            },
+            {
+                "uri": "http://example.org/Manager",
+                "name": "Manager",
+                "subClassOf": "http://example.org/Employee",
+                "equivalentClass": "http://example.org/Supervisor",
+            },
+        ],
+        "object_properties": [
+            {
+                "uri": "http://example.org/worksFor",
+                "name": "worksFor",
+                "domain": "http://example.org/Employee",
+                "range": "http://example.org/Company",
+            },
+            {
+                "uri": "http://example.org/manages",
+                "name": "manages",
+                "comment": "manages a team",
+                "domain": ["http://example.org/Manager"],
+                "range": ["http://example.org/Employee"],
+            },
+        ],
+        "data_properties": [
+            {
+                "uri": "http://example.org/hasAge",
+                "name": "hasAge",
+                "domain": "http://example.org/Person",
+                "range": "integer",
+            },
+            {
+                "uri": "http://example.org/hasName",
+                "name": "hasName",
+                "comment": "full name",
+                "domain": "http://example.org/Person",
+                "range": "string",
+            },
+        ],
+    }
+
+
+# ---------------------------------------------------------------------------
+# Bug 1 — valid Turtle syntax
+# ---------------------------------------------------------------------------
+
+class TestTurtleSyntaxValidity:
+    """Every subject block must have exactly one closing period at the end."""
+
+    def _blocks(self, turtle: str) -> list[str]:
+        """Split output into non-empty logical blocks (separated by blank lines)."""
+        return [b.strip() for b in turtle.split("\n\n") if b.strip()]
+
+    def test_no_triple_after_period(self, exporter, full_ontology):
+        """No predicate line may appear after a line that ends with ' .'."""
+        turtle = exporter._export_owl_turtle(full_ontology)
+        lines = turtle.splitlines()
+        for i, line in enumerate(lines):
+            stripped = line.rstrip()
+            if stripped.endswith(" .") and i + 1 < len(lines):
+                next_line = lines[i + 1].strip()
+                # next non-blank line must not be a predicate continuation
+                if next_line:
+                    assert not next_line.startswith("rdfs:"), (
+                        f"Predicate continuation after closing '.' at line {i + 1}: "
+                        f"{lines[i]!r} → {lines[i + 1]!r}"
+                    )
+
+    def test_each_subject_block_ends_with_period(self, exporter, full_ontology):
+        """Every subject block (class / property declaration) ends with exactly one '.'."""
+        turtle = exporter._export_owl_turtle(full_ontology)
+        blocks = self._blocks(turtle)
+        # skip the @prefix lines block and ontology declaration
+        subject_blocks = [b for b in blocks if b.startswith("<http://")]
+        for block in subject_blocks:
+            assert block.endswith("."), f"Block does not end with '.': {block!r}"
+            # Must not have a bare '.' on an interior line
+            interior_lines = block.splitlines()[:-1]
+            for ln in interior_lines:
+                assert not ln.rstrip().endswith(" ."), (
+                    f"Premature closing period inside block: {ln!r}"
+                )
+
+    def test_class_with_subclassof_is_valid(self, exporter):
+        ontology = {
+            "uri": "http://example.org/onto",
+            "name": "T",
+            "classes": [
+                {
+                    "uri": "http://example.org/Employee",
+                    "name": "Employee",
+                    "subClassOf": "http://example.org/Person",
+                }
+            ],
+            "object_properties": [],
+            "data_properties": [],
+        }
+        turtle = exporter._export_owl_turtle(ontology)
+        # Must contain both predicates in the same block
+        assert 'rdfs:label "Employee"' in turtle
+        assert "rdfs:subClassOf <http://example.org/Person>" in turtle
+        # The subClassOf line must NOT come after a closing period
+        lines = turtle.splitlines()
+        for i, ln in enumerate(lines):
+            if "rdfs:subClassOf" in ln:
+                # Search backwards for the closest period-terminated line
+                for prev in reversed(lines[:i]):
+                    prev_s = prev.rstrip()
+                    if prev_s:
+                        assert not prev_s.endswith(" ."), (
+                            "rdfs:subClassOf appeared after a closed block"
+                        )
+                        break
+
+    def test_object_property_with_domain_range_is_valid(self, exporter):
+        ontology = {
+            "uri": "http://example.org/onto",
+            "name": "T",
+            "classes": [],
+            "object_properties": [
+                {
+                    "uri": "http://example.org/worksFor",
+                    "name": "worksFor",
+                    "domain": "http://example.org/Employee",
+                    "range": "http://example.org/Company",
+                }
+            ],
+            "data_properties": [],
+        }
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "rdfs:domain <http://example.org/Employee>" in turtle
+        assert "rdfs:range <http://example.org/Company>" in turtle
+        lines = turtle.splitlines()
+        for i, ln in enumerate(lines):
+            if "rdfs:domain" in ln or "rdfs:range" in ln:
+                for prev in reversed(lines[:i]):
+                    prev_s = prev.rstrip()
+                    if prev_s:
+                        assert not prev_s.endswith(" ."), (
+                            "domain/range appeared after a closed block"
+                        )
+                        break
+
+    def test_class_with_comment_subclassof_both_present(self, exporter):
+        ontology = {
+            "uri": "http://example.org/onto",
+            "name": "T",
+            "classes": [
+                {
+                    "uri": "http://example.org/X",
+                    "name": "X",
+                    "comment": "some comment",
+                    "subClassOf": "http://example.org/Y",
+                }
+            ],
+            "object_properties": [],
+            "data_properties": [],
+        }
+        turtle = exporter._export_owl_turtle(ontology)
+        assert 'rdfs:comment "some comment"' in turtle
+        assert "rdfs:subClassOf <http://example.org/Y>" in turtle
+        # block must end with single period
+        block = [b for b in turtle.split("\n\n") if "owl:Class" in b][0].strip()
+        assert block.endswith(".")
+        assert block.count("\n.") == 0  # no bare period-only lines
+
+
+# ---------------------------------------------------------------------------
+# Bug 2 — data properties present in Turtle output
+# ---------------------------------------------------------------------------
+
+class TestDataPropertiesInTurtle:
+
+    def test_data_property_declared_as_datatypeproperty(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        assert "owl:DatatypeProperty" in turtle
+
+    def test_data_property_uri_present(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        assert "<http://example.org/hasAge>" in turtle
+        assert "<http://example.org/hasName>" in turtle
+
+    def test_data_property_label(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        assert 'rdfs:label "hasAge"' in turtle
+        assert 'rdfs:label "hasName"' in turtle
+
+    def test_data_property_domain(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        assert "rdfs:domain <http://example.org/Person>" in turtle
+
+    def test_data_property_range_uses_xsd_prefix(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        assert "rdfs:range xsd:integer" in turtle
+        assert "rdfs:range xsd:string" in turtle
+
+    def test_data_property_comment(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        assert 'rdfs:comment "full name"' in turtle
+
+    def test_data_properties_not_in_turtle_was_bug(self, exporter):
+        """Regression: data_properties were silently dropped before the fix."""
+        ontology = {
+            "uri": "http://example.org/onto",
+            "name": "T",
+            "classes": [],
+            "object_properties": [],
+            "data_properties": [
+                {
+                    "uri": "http://example.org/birthDate",
+                    "name": "birthDate",
+                    "range": "date",
+                }
+            ],
+        }
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "owl:DatatypeProperty" in turtle, (
+            "Data properties must appear in Turtle output (was silently dropped)"
+        )
+        assert "<http://example.org/birthDate>" in turtle
+        assert "rdfs:range xsd:date" in turtle
+
+    def test_data_property_block_ends_with_period(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        blocks = [b.strip() for b in turtle.split("\n\n") if "owl:DatatypeProperty" in b]
+        assert blocks, "Expected at least one DatatypeProperty block"
+        for block in blocks:
+            assert block.endswith("."), f"DatatypeProperty block missing closing '.': {block!r}"
+
+
+# ---------------------------------------------------------------------------
+# Namespace and ontology header
+# ---------------------------------------------------------------------------
+
+class TestTurtleHeader:
+
+    def test_prefix_declarations(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        assert "@prefix rdf:" in turtle
+        assert "@prefix rdfs:" in turtle
+        assert "@prefix owl:" in turtle
+        assert "@prefix xsd:" in turtle
+
+    def test_ontology_declaration(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        assert "a owl:Ontology" in turtle
+        assert 'rdfs:label "TestOntology"' in turtle
+        assert 'owl:versionInfo "1.0"' in turtle
+
+    def test_ontology_description_included(self, exporter, full_ontology):
+        turtle = exporter._export_owl_turtle(full_ontology)
+        assert 'rdfs:comment "A test ontology"' in turtle
+
+    def test_ontology_without_description(self, exporter):
+        ontology = {"uri": "http://example.org/onto", "name": "NoDesc",
+                    "classes": [], "object_properties": [], "data_properties": []}
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "rdfs:comment" not in turtle
+
+
+# ---------------------------------------------------------------------------
+# Object properties — list domain/range
+# ---------------------------------------------------------------------------
+
+class TestObjectPropertyListDomainRange:
+
+    def test_list_domain(self, exporter):
+        ontology = {
+            "uri": "http://example.org/onto", "name": "T",
+            "classes": [],
+            "object_properties": [
+                {
+                    "uri": "http://example.org/p",
+                    "name": "p",
+                    "domain": ["http://example.org/A", "http://example.org/B"],
+                }
+            ],
+            "data_properties": [],
+        }
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "rdfs:domain <http://example.org/A>" in turtle
+        assert "rdfs:domain <http://example.org/B>" in turtle
+
+    def test_list_range(self, exporter):
+        ontology = {
+            "uri": "http://example.org/onto", "name": "T",
+            "classes": [],
+            "object_properties": [
+                {
+                    "uri": "http://example.org/p",
+                    "name": "p",
+                    "range": ["http://example.org/X", "http://example.org/Y"],
+                }
+            ],
+            "data_properties": [],
+        }
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "rdfs:range <http://example.org/X>" in turtle
+        assert "rdfs:range <http://example.org/Y>" in turtle
+
+
+# ---------------------------------------------------------------------------
+# equivalentClass support (also tested under Bug 1 guard)
+# ---------------------------------------------------------------------------
+
+class TestEquivalentClass:
+
+    def test_equivalent_class_in_turtle(self, exporter):
+        ontology = {
+            "uri": "http://example.org/onto", "name": "T",
+            "classes": [
+                {
+                    "uri": "http://example.org/Manager",
+                    "name": "Manager",
+                    "equivalentClass": "http://example.org/Supervisor",
+                }
+            ],
+            "object_properties": [],
+            "data_properties": [],
+        }
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "owl:equivalentClass <http://example.org/Supervisor>" in turtle
+        block = [b for b in turtle.split("\n\n") if "owl:Class" in b][0].strip()
+        assert block.endswith(".")
+
+
+# ---------------------------------------------------------------------------
+# String escaping in Turtle literals (issue #478 review — escape_001)
+# ---------------------------------------------------------------------------
+
+class TestTurtleStringEscaping:
+    """User-provided strings must be escaped before embedding in Turtle literals."""
+
+    def _onto(self, **kwargs):
+        base = {"uri": "http://example.org/onto", "name": "T",
+                "classes": [], "object_properties": [], "data_properties": []}
+        base.update(kwargs)
+        return base
+
+    def test_escape_ttl_str_double_quote(self, exporter):
+        assert exporter._escape_ttl_str('say "hello"') == r'say \"hello\"'
+
+    def test_escape_ttl_str_backslash(self, exporter):
+        assert exporter._escape_ttl_str("C:\\path") == "C:\\\\path"
+
+    def test_escape_ttl_str_newline(self, exporter):
+        assert exporter._escape_ttl_str("line1\nline2") == "line1\\nline2"
+
+    def test_escape_ttl_str_carriage_return(self, exporter):
+        assert exporter._escape_ttl_str("a\rb") == "a\\rb"
+
+    def test_escape_ttl_str_tab(self, exporter):
+        assert exporter._escape_ttl_str("col1\tcol2") == "col1\\tcol2"
+
+    def test_escape_ttl_str_combined(self, exporter):
+        raw = 'back\\slash and "quote"\nnewline'
+        escaped = exporter._escape_ttl_str(raw)
+        assert '\\"' in escaped
+        assert "\\\\" in escaped
+        assert "\\n" in escaped
+
+    def test_ontology_name_with_quote_is_escaped(self, exporter):
+        ontology = self._onto(name='John"s Ontology')
+        turtle = exporter._export_owl_turtle(ontology)
+        assert 'rdfs:label "John\\"s Ontology"' in turtle
+        assert 'rdfs:label "John"s Ontology"' not in turtle
+
+    def test_ontology_description_with_quote_is_escaped(self, exporter):
+        ontology = self._onto(description='Describes "things"')
+        turtle = exporter._export_owl_turtle(ontology)
+        assert 'rdfs:comment "Describes \\"things\\""' in turtle
+
+    def test_class_name_with_quote_is_escaped(self, exporter):
+        ontology = self._onto(classes=[{
+            "uri": "http://example.org/C",
+            "name": 'My "Special" Class',
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert r'rdfs:label "My \"Special\" Class"' in turtle
+
+    def test_class_comment_with_backslash_is_escaped(self, exporter):
+        ontology = self._onto(classes=[{
+            "uri": "http://example.org/C",
+            "name": "C",
+            "comment": "path is C:\\Users",
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert r'rdfs:comment "path is C:\\Users"' in turtle
+
+    def test_class_comment_with_newline_is_escaped(self, exporter):
+        ontology = self._onto(classes=[{
+            "uri": "http://example.org/C",
+            "name": "C",
+            "comment": "line1\nline2",
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert r'rdfs:comment "line1\nline2"' in turtle
+
+    def test_object_property_name_with_quote_is_escaped(self, exporter):
+        ontology = self._onto(object_properties=[{
+            "uri": "http://example.org/p",
+            "name": 'has"Value',
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert r'rdfs:label "has\"Value"' in turtle
+
+    def test_object_property_comment_with_quote_is_escaped(self, exporter):
+        ontology = self._onto(object_properties=[{
+            "uri": "http://example.org/p",
+            "name": "p",
+            "comment": 'links "A" to "B"',
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert r'rdfs:comment "links \"A\" to \"B\""' in turtle
+
+    def test_data_property_name_with_quote_is_escaped(self, exporter):
+        ontology = self._onto(data_properties=[{
+            "uri": "http://example.org/dp",
+            "name": 'the "name" prop',
+            "range": "string",
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert r'rdfs:label "the \"name\" prop"' in turtle
+
+    def test_data_property_comment_with_quote_is_escaped(self, exporter):
+        ontology = self._onto(data_properties=[{
+            "uri": "http://example.org/dp",
+            "name": "dp",
+            "comment": 'see "spec" §3',
+            "range": "string",
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert r'rdfs:comment "see \"spec\" §3"' in turtle
+
+    def test_plain_strings_unchanged(self, exporter):
+        """Strings without special chars must pass through unchanged."""
+        ontology = self._onto(
+            name="MyOntology",
+            classes=[{"uri": "http://example.org/C", "name": "SafeName"}],
+        )
+        turtle = exporter._export_owl_turtle(ontology)
+        assert 'rdfs:label "MyOntology"' in turtle
+        assert 'rdfs:label "SafeName"' in turtle
+
+
+# ---------------------------------------------------------------------------
+# Null / missing optional fields — no KeyError raised (review null_check_001-3)
+# ---------------------------------------------------------------------------
+
+class TestNullFieldHandling:
+    """Optional fields absent from dicts must not raise KeyError."""
+
+    def _onto(self, **kwargs):
+        base = {"uri": "http://example.org/onto", "name": "T",
+                "classes": [], "object_properties": [], "data_properties": []}
+        base.update(kwargs)
+        return base
+
+    def test_class_no_optional_fields(self, exporter):
+        ontology = self._onto(classes=[{"uri": "http://example.org/C", "name": "C"}])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "owl:Class" in turtle
+
+    def test_object_property_no_domain_no_range(self, exporter):
+        ontology = self._onto(object_properties=[{
+            "uri": "http://example.org/p", "name": "p"
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "owl:ObjectProperty" in turtle
+        assert "rdfs:domain" not in turtle
+        assert "rdfs:range" not in turtle
+
+    def test_data_property_no_domain_no_range(self, exporter):
+        ontology = self._onto(data_properties=[{
+            "uri": "http://example.org/dp", "name": "dp"
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "owl:DatatypeProperty" in turtle
+        assert "rdfs:domain" not in turtle
+        assert "rdfs:range" not in turtle
+
+    def test_data_property_none_domain(self, exporter):
+        """Explicit None value for domain must not raise KeyError."""
+        ontology = self._onto(data_properties=[{
+            "uri": "http://example.org/dp", "name": "dp",
+            "domain": None, "range": "string",
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "rdfs:domain" not in turtle
+
+    def test_data_property_none_range(self, exporter):
+        """Explicit None value for range must not raise KeyError."""
+        ontology = self._onto(data_properties=[{
+            "uri": "http://example.org/dp", "name": "dp",
+            "domain": "http://example.org/C", "range": None,
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "rdfs:range" not in turtle
+
+    def test_object_property_none_domain(self, exporter):
+        ontology = self._onto(object_properties=[{
+            "uri": "http://example.org/p", "name": "p",
+            "domain": None, "range": "http://example.org/X",
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "rdfs:domain" not in turtle
+
+    def test_object_property_none_range(self, exporter):
+        ontology = self._onto(object_properties=[{
+            "uri": "http://example.org/p", "name": "p",
+            "domain": "http://example.org/A", "range": None,
+        }])
+        turtle = exporter._export_owl_turtle(ontology)
+        assert "rdfs:range" not in turtle

tests/kg/test_path_finder.py

@@ -821,3 +821,87 @@ class TestPathFinderEdgeCases:
         
         paths = self.finder.all_shortest_paths(single_node_graph, "A")
         assert len(paths) == 0  # No paths to other nodes
+
+
+class TestBidirectionalPathFinding:
+    """Tests for the directed=False undirected-traversal mode (issue #469)."""
+
+    def setup_method(self):
+        self.finder = PathFinder()
+        # Single directed edge A → B.  Reverse query B → A has no directed path.
+        self.digraph = nx.DiGraph()
+        self.digraph.add_edge("A", "B")
+
+    # --- directed=True (default) preserves existing behaviour ---
+
+    def test_bfs_directed_true_reverse_returns_empty(self):
+        """B→A should find nothing when directed=True (default)."""
+        path = self.finder.bfs_shortest_path(self.digraph, "B", "A", directed=True)
+        assert path == []
+
+    def test_dijkstra_directed_true_reverse_returns_empty(self):
+        """B→A should find nothing when directed=True (default)."""
+        path = self.finder.dijkstra_shortest_path(self.digraph, "B", "A", directed=True)
+        assert path == []
+
+    def test_bfs_directed_true_default_arg(self):
+        """Omitting directed= should behave the same as directed=True."""
+        path = self.finder.bfs_shortest_path(self.digraph, "B", "A")
+        assert path == []
+
+    def test_dijkstra_directed_true_default_arg(self):
+        path = self.finder.dijkstra_shortest_path(self.digraph, "B", "A")
+        assert path == []
+
+    # --- directed=False finds path against edge orientation ---
+
+    def test_bfs_directed_false_reverse_single_edge(self):
+        """directed=False must find B→A even though only A→B exists."""
+        path = self.finder.bfs_shortest_path(self.digraph, "B", "A", directed=False)
+        assert path == ["B", "A"]
+
+    def test_dijkstra_directed_false_reverse_single_edge(self):
+        path = self.finder.dijkstra_shortest_path(self.digraph, "B", "A", directed=False)
+        assert path == ["B", "A"]
+
+    def test_bfs_directed_false_forward_still_works(self):
+        """directed=False should not break the forward direction."""
+        path = self.finder.bfs_shortest_path(self.digraph, "A", "B", directed=False)
+        assert path == ["A", "B"]
+
+    def test_dijkstra_directed_false_forward_still_works(self):
+        path = self.finder.dijkstra_shortest_path(self.digraph, "A", "B", directed=False)
+        assert path == ["A", "B"]
+
+    # --- multi-hop path where one edge is against the query direction ---
+
+    def test_bfs_directed_false_multihop(self):
+        """A→B, C→B graph: directed=False lets us find A→B→C (i.e. A→C via B)."""
+        g = nx.DiGraph()
+        g.add_edge("A", "B")
+        g.add_edge("C", "B")  # oriented towards B, not away from it
+        # undirected view: A-B-C, so A→C path exists
+        path = self.finder.bfs_shortest_path(g, "A", "C", directed=False)
+        assert path[0] == "A" and path[-1] == "C"
+        assert "B" in path
+
+    def test_dijkstra_directed_false_multihop(self):
+        g = nx.DiGraph()
+        g.add_edge("A", "B")
+        g.add_edge("C", "B")
+        path = self.finder.dijkstra_shortest_path(g, "A", "C", directed=False)
+        assert path[0] == "A" and path[-1] == "C"
+        assert "B" in path
+
+    # --- PathResponse.directed field ---
+
+    def test_path_response_directed_field_exists(self):
+        """PathResponse must carry a directed field."""
+        from semantica.explorer.schemas import PathResponse
+        r = PathResponse(source="A", target="B", algorithm="bfs", path=["A", "B"], directed=False)
+        assert r.directed is False
+
+    def test_path_response_directed_field_defaults_true(self):
+        from semantica.explorer.schemas import PathResponse
+        r = PathResponse(source="A", target="B", algorithm="bfs", path=["A", "B"])
+        assert r.directed is True

tests/semantic_extract/test_pr482_deepseek_openai.py

@@ -0,0 +1,348 @@
+"""Tests for PR #482: DeepSeekProvider switch from deepseek SDK to openai SDK."""
+
+import sys
+import os
+import unittest
+from unittest.mock import patch, MagicMock, call
+from pydantic import BaseModel
+
+sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "../../")))
+
+
+class TestDeepSeekProviderInit(unittest.TestCase):
+    """Tests for DeepSeekProvider.__init__ and _init_client after PR #482."""
+
+    def setUp(self):
+        from semantica.semantic_extract.providers import DeepSeekProvider
+        self.DeepSeekProvider = DeepSeekProvider
+
+    def test_base_url_set_on_init(self):
+        """self.base_url must be set before _init_client is called (PR #482 regression)."""
+        with patch.object(self.DeepSeekProvider, "_init_client", return_value=None):
+            provider = self.DeepSeekProvider(api_key="fake-key")
+        self.assertTrue(
+            hasattr(provider, "base_url"),
+            "DeepSeekProvider missing self.base_url — causes AttributeError in _init_client",
+        )
+        self.assertEqual(provider.base_url, "https://api.deepseek.com/v1")
+
+    def test_base_url_points_to_v1_endpoint(self):
+        """base_url must include /v1 so OpenAI SDK resolves /chat/completions correctly."""
+        with patch.object(self.DeepSeekProvider, "_init_client", return_value=None):
+            provider = self.DeepSeekProvider(api_key="fake-key")
+        self.assertIn("/v1", provider.base_url, "base_url must include /v1")
+
+    def test_init_client_uses_openai_not_deepseek(self):
+        """_init_client must import openai.OpenAI, not deepseek.Client."""
+        mock_openai_cls = MagicMock()
+        mock_openai_instance = MagicMock()
+        mock_openai_cls.return_value = mock_openai_instance
+
+        with patch.dict("sys.modules", {"openai": MagicMock(OpenAI=mock_openai_cls)}):
+            # Re-import to pick up patched sys.modules
+            import importlib
+            import semantica.semantic_extract.providers as providers_mod
+            importlib.reload(providers_mod)
+            DeepSeekProvider = providers_mod.DeepSeekProvider
+
+            provider = DeepSeekProvider(api_key="sk-test")
+
+        mock_openai_cls.assert_called_once_with(
+            api_key="sk-test",
+            base_url="https://api.deepseek.com/v1",
+        )
+        self.assertIs(provider.client, mock_openai_instance)
+
+    def test_init_client_no_api_key_leaves_client_none(self):
+        """Without an API key, client must remain None."""
+        with patch("semantica.semantic_extract.providers.config") as mock_cfg:
+            mock_cfg.get_api_key.return_value = None
+            with patch.object(self.DeepSeekProvider, "_init_client", return_value=None):
+                provider = self.DeepSeekProvider(api_key=None)
+            provider.client = None  # simulate _init_client no-op
+        self.assertFalse(provider.is_available())
+
+    def test_init_client_handles_openai_import_error(self):
+        """If openai is not installed, _init_client must set client=None, not raise."""
+        with patch.object(self.DeepSeekProvider, "_init_client", return_value=None):
+            provider = self.DeepSeekProvider(api_key="sk-test")
+        provider.client = None  # manually simulate ImportError path
+        # Directly call _init_client with openai blocked
+        with patch.dict("sys.modules", {"openai": None}):
+            try:
+                provider._init_client()
+            except Exception as e:
+                self.fail(f"_init_client raised unexpectedly: {e}")
+        self.assertIsNone(provider.client)
+
+    def test_is_available_true_when_client_set(self):
+        """is_available() returns True when self.client is an OpenAI instance."""
+        with patch.object(self.DeepSeekProvider, "_init_client", return_value=None):
+            provider = self.DeepSeekProvider(api_key="sk-test")
+        provider.client = MagicMock()
+        self.assertTrue(provider.is_available())
+
+    def test_is_available_false_when_client_none(self):
+        """is_available() returns False when self.client is None."""
+        with patch.object(self.DeepSeekProvider, "_init_client", return_value=None):
+            provider = self.DeepSeekProvider(api_key="sk-test")
+        provider.client = None
+        self.assertFalse(provider.is_available())
+
+    def test_no_deepseek_module_imported(self):
+        """deepseek module must NOT be imported by _init_client after PR #482."""
+        with patch.object(self.DeepSeekProvider, "_init_client", return_value=None):
+            provider = self.DeepSeekProvider(api_key="sk-test")
+        provider.client = None
+
+        blocked = MagicMock()
+        blocked.__spec__ = None
+        with patch.dict("sys.modules", {"deepseek": None}):
+            # _init_client should succeed even if deepseek is completely absent
+            mock_openai = MagicMock()
+            mock_openai.OpenAI.return_value = MagicMock()
+            with patch.dict("sys.modules", {"openai": mock_openai, "deepseek": None}):
+                try:
+                    provider._init_client()
+                except Exception as e:
+                    self.fail(f"_init_client raised when deepseek absent: {e}")
+
+
+class TestDeepSeekProviderGenerate(unittest.TestCase):
+    """Tests for DeepSeekProvider.generate / generate_structured with OpenAI client."""
+
+    def _make_provider(self, api_key="sk-test"):
+        from semantica.semantic_extract.providers import DeepSeekProvider
+        with patch.object(DeepSeekProvider, "_init_client", return_value=None):
+            provider = DeepSeekProvider(api_key=api_key)
+        provider.client = MagicMock()
+        return provider
+
+    def test_generate_uses_chat_completions(self):
+        """generate() must call client.chat.completions.create."""
+        provider = self._make_provider()
+        mock_resp = MagicMock()
+        mock_resp.choices[0].message.content = "hello"
+        provider.client.chat.completions.create.return_value = mock_resp
+
+        result = provider.generate("test prompt")
+
+        provider.client.chat.completions.create.assert_called_once()
+        self.assertEqual(result, "hello")
+
+    def test_generate_passes_model(self):
+        provider = self._make_provider()
+        mock_resp = MagicMock()
+        mock_resp.choices[0].message.content = "x"
+        provider.client.chat.completions.create.return_value = mock_resp
+
+        provider.generate("p", model="deepseek-reasoner")
+        kwargs = provider.client.chat.completions.create.call_args[1]
+        self.assertEqual(kwargs["model"], "deepseek-reasoner")
+
+    def test_generate_structured_returns_parsed_json(self):
+        provider = self._make_provider()
+        mock_resp = MagicMock()
+        mock_resp.choices[0].message.content = '{"key": "value"}'
+        provider.client.chat.completions.create.return_value = mock_resp
+
+        result = provider.generate_structured("test prompt")
+        self.assertEqual(result, {"key": "value"})
+
+    def test_generate_raises_without_client(self):
+        from semantica.semantic_extract.providers import DeepSeekProvider, ProcessingError
+        with patch.object(DeepSeekProvider, "_init_client", return_value=None):
+            provider = DeepSeekProvider(api_key="sk-test")
+        provider.client = None
+
+        with self.assertRaises(ProcessingError):
+            provider.generate("prompt")
+
+    def test_generate_structured_raises_without_client(self):
+        from semantica.semantic_extract.providers import DeepSeekProvider, ProcessingError
+        with patch.object(DeepSeekProvider, "_init_client", return_value=None):
+            provider = DeepSeekProvider(api_key="sk-test")
+        provider.client = None
+
+        with self.assertRaises(ProcessingError):
+            provider.generate_structured("prompt")
+
+
+class TestDeepSeekInstructorPath(unittest.TestCase):
+    """Tests for generate_typed instructor path with DeepSeekProvider (OpenAI client)."""
+
+    def _make_provider(self, api_key="sk-test"):
+        from semantica.semantic_extract.providers import DeepSeekProvider
+        from unittest.mock import MagicMock
+        from openai import OpenAI
+        with patch.object(DeepSeekProvider, "_init_client", return_value=None):
+            provider = DeepSeekProvider(api_key=api_key)
+        # After PR #482, client is an OpenAI instance
+        mock_client = MagicMock(spec=OpenAI)
+        provider.client = mock_client
+        return provider
+
+    def test_generate_typed_instructor_openai_isinstance_check(self):
+        """After PR #482, client is OpenAI, so instructor path must use from_openai."""
+        from openai import OpenAI
+        from semantica.semantic_extract.providers import DeepSeekProvider
+        with patch.object(DeepSeekProvider, "_init_client", return_value=None):
+            provider = DeepSeekProvider(api_key="sk-test")
+        provider.client = MagicMock(spec=OpenAI)
+
+        self.assertIsInstance(
+            provider.client, OpenAI,
+            "client must be OpenAI instance for instructor isinstance check to pass",
+        )
+
+
+class TestVerboseModeAssignment(unittest.TestCase):
+    """Tests for verbose_mode assignment fix in BaseProvider.generate_typed (commit eec3e88)."""
+
+    def _make_openai_provider(self):
+        from semantica.semantic_extract.providers import OpenAIProvider
+        with patch.object(OpenAIProvider, "_init_client", return_value=None):
+            provider = OpenAIProvider(api_key="sk-test")
+        provider.client = MagicMock()
+        return provider
+
+    def test_generate_typed_no_verbose_no_name_error(self):
+        """generate_typed must not raise NameError for verbose_mode when verbose not passed."""
+        provider = self._make_openai_provider()
+
+        class Schema(BaseModel):
+            value: str
+
+        mock_instructor = MagicMock()
+        mock_client = MagicMock()
+        mock_client.chat.completions.create.return_value = Schema(value="ok")
+        mock_instructor.from_openai.return_value = mock_client
+        mock_instructor.from_provider.side_effect = Exception("skip")
+        mock_instructor.Mode.TOOLS = "tools"
+
+        with patch("semantica.semantic_extract.providers.instructor", mock_instructor):
+            try:
+                result = provider.generate_typed("prompt", Schema)
+            except NameError as e:
+                self.fail(f"NameError for verbose_mode: {e}")
+            except Exception:
+                pass  # other errors are OK — we only care NameError is gone
+
+    def test_generate_typed_verbose_true_prints(self):
+        """When verbose=True, generate_typed must print the confirmation line."""
+        provider = self._make_openai_provider()
+
+        class Schema(BaseModel):
+            value: str
+
+        mock_schema_instance = Schema(value="ok")
+        mock_instructor = MagicMock()
+        mock_ic_client = MagicMock()
+        mock_ic_client.chat.completions.create.return_value = mock_schema_instance
+        mock_instructor.from_openai.return_value = mock_ic_client
+        mock_instructor.from_provider.side_effect = Exception("skip")
+        mock_instructor.Mode.TOOLS = "tools"
+
+        import io
+        captured = io.StringIO()
+        with patch("semantica.semantic_extract.providers.instructor", mock_instructor):
+            with patch("sys.stdout", captured):
+                try:
+                    provider.generate_typed("prompt", Schema, verbose=True)
+                except Exception:
+                    pass
+
+        output = captured.getvalue()
+        # verbose_mode=True should trigger the print statement
+        self.assertIn("generate_typed", output)
+
+    def test_generate_typed_verbose_false_no_print(self):
+        """When verbose=False (default), generate_typed must not print anything."""
+        provider = self._make_openai_provider()
+
+        class Schema(BaseModel):
+            value: str
+
+        mock_schema_instance = Schema(value="ok")
+        mock_instructor = MagicMock()
+        mock_ic_client = MagicMock()
+        mock_ic_client.chat.completions.create.return_value = mock_schema_instance
+        mock_instructor.from_openai.return_value = mock_ic_client
+        mock_instructor.from_provider.side_effect = Exception("skip")
+        mock_instructor.Mode.TOOLS = "tools"
+
+        import io
+        captured = io.StringIO()
+        with patch("semantica.semantic_extract.providers.instructor", mock_instructor):
+            with patch("sys.stdout", captured):
+                try:
+                    provider.generate_typed("prompt", Schema)
+                except Exception:
+                    pass
+
+        self.assertEqual(captured.getvalue(), "")
+
+    def test_generate_typed_verbose_from_config(self):
+        """verbose_mode must also respect config-level verbose setting."""
+        provider = self._make_openai_provider()
+        provider.config["verbose"] = True
+
+        class Schema(BaseModel):
+            value: str
+
+        mock_schema_instance = Schema(value="ok")
+        mock_instructor = MagicMock()
+        mock_ic_client = MagicMock()
+        mock_ic_client.chat.completions.create.return_value = mock_schema_instance
+        mock_instructor.from_openai.return_value = mock_ic_client
+        mock_instructor.from_provider.side_effect = Exception("skip")
+        mock_instructor.Mode.TOOLS = "tools"
+
+        import io
+        captured = io.StringIO()
+        with patch("semantica.semantic_extract.providers.instructor", mock_instructor):
+            with patch("sys.stdout", captured):
+                try:
+                    provider.generate_typed("prompt", Schema)
+                except Exception:
+                    pass
+
+        self.assertIn("generate_typed", captured.getvalue())
+
+
+class TestDeepSeekGenerateTypedInstructorIntegration(unittest.TestCase):
+    """Integration-style tests: DeepSeekProvider.generate_typed with instructor."""
+
+    def test_generate_typed_deepseek_uses_openai_client_for_instructor(self):
+        """generate_typed instructor path for DeepSeek must reuse the OpenAI client."""
+        from semantica.semantic_extract.providers import DeepSeekProvider
+        from openai import OpenAI
+
+        with patch.object(DeepSeekProvider, "_init_client", return_value=None):
+            provider = DeepSeekProvider(api_key="sk-test")
+        mock_openai_client = MagicMock(spec=OpenAI)
+        provider.client = mock_openai_client
+
+        class Schema(BaseModel):
+            label: str
+
+        mock_instructor = MagicMock()
+        mock_ic_client = MagicMock()
+        mock_ic_client.chat.completions.create.return_value = Schema(label="ok")
+        mock_instructor.from_openai.return_value = mock_ic_client
+        mock_instructor.from_provider.side_effect = Exception("no from_provider")
+        mock_instructor.Mode.JSON = "json"
+        mock_instructor.Mode.TOOLS = "tools"
+
+        with patch("semantica.semantic_extract.providers.instructor", mock_instructor):
+            result = provider.generate_typed("classify this", Schema)
+
+        # Must have called from_openai with the existing client (not a fresh one)
+        mock_instructor.from_openai.assert_called_once_with(
+            mock_openai_client, mode="json"
+        )
+        self.assertEqual(result.label, "ok")
+
+
+if __name__ == "__main__":
+    unittest.main()

tests/visualization/test_kg_visualizer_normalize_graph.py

@@ -283,5 +283,166 @@ class TestAllVisualizeMethodsAcceptKGObject(unittest.TestCase):
         self.viz._normalize_graph.assert_called_once_with(self.kg)
 
 
+# ---------------------------------------------------------------------------
+# Issue #471 — formal KnowledgeGraph type support
+# ---------------------------------------------------------------------------
+
+class TestFormalKnowledgeGraphType(unittest.TestCase):
+    """
+    Regression tests for issue #471.
+
+    The formal ``semantica.kg.KnowledgeGraph`` dataclass must be accepted by
+    every public visualize_* method without requiring any manual conversion.
+    """
+
+    @classmethod
+    def setUpClass(cls):
+        try:
+            from semantica.kg.knowledge_graph import KnowledgeGraph
+            cls.KnowledgeGraph = KnowledgeGraph
+        except ImportError:
+            cls.KnowledgeGraph = None
+
+    def _make_kg(self):
+        if self.KnowledgeGraph is None:
+            self.skipTest("semantica.kg.KnowledgeGraph not available")
+        return self.KnowledgeGraph(
+            entities=ENTITIES,
+            relationships=RELATIONSHIPS,
+            metadata={"version": "test"},
+        )
+
+    def test_convert_knowledge_graph_entities(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        result = viz._convert_knowledge_graph(kg)
+        self.assertEqual(result["entities"], ENTITIES)
+
+    def test_convert_knowledge_graph_relationships(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        result = viz._convert_knowledge_graph(kg)
+        self.assertEqual(result["relationships"], RELATIONSHIPS)
+
+    def test_convert_knowledge_graph_metadata(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        result = viz._convert_knowledge_graph(kg)
+        self.assertEqual(result["metadata"], {"version": "test"})
+
+    def test_convert_knowledge_graph_does_not_mutate(self):
+        kg = self._make_kg()
+        original_entities = list(kg.entities)
+        original_relationships = list(kg.relationships)
+        viz = _make_viz()
+        viz._convert_knowledge_graph(kg)
+        self.assertEqual(kg.entities, original_entities)
+        self.assertEqual(kg.relationships, original_relationships)
+
+    def test_convert_knowledge_graph_is_deterministic(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        self.assertEqual(viz._convert_knowledge_graph(kg), viz._convert_knowledge_graph(kg))
+
+    def test_normalize_graph_routes_kg_type(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        viz._convert_knowledge_graph = MagicMock(return_value=GRAPH_DICT)
+        viz._normalize_graph(kg)
+        viz._convert_knowledge_graph.assert_called_once_with(kg)
+
+    def test_normalize_graph_returns_dict_for_kg_type(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        result = viz._normalize_graph(kg)
+        self.assertIsInstance(result, dict)
+        self.assertIn("entities", result)
+        self.assertIn("relationships", result)
+
+    def _run_visualize_network(self, graph_arg):
+        mock_fig = MagicMock()
+        mock_go = sys.modules["plotly.graph_objects"]
+        mock_go.Figure.return_value = mock_fig
+        mock_go.Scatter.return_value = MagicMock()
+        mock_go.Layout.return_value = MagicMock()
+        viz = _make_viz()
+        fake_pos = {"e1": (0.0, 0.0), "e2": (1.0, 1.0)}
+        viz.force_layout = MagicMock()
+        viz.force_layout.compute_layout.return_value = fake_pos
+        viz.hierarchical_layout = MagicMock()
+        viz.circular_layout = MagicMock()
+        with (
+            patch(
+                "semantica.visualization.kg_visualizer.ColorPalette.get_entity_type_colors",
+                return_value={"Person": "#ff0000"},
+            ),
+            patch(
+                "semantica.visualization.kg_visualizer.ColorPalette.get_colors",
+                return_value=["#ff0000"],
+            ),
+        ):
+            return viz.visualize_network(graph_arg, output="interactive")
+
+    def test_visualize_network_accepts_knowledge_graph(self):
+        self.assertIsNotNone(self._run_visualize_network(self._make_kg()))
+
+    def test_visualize_communities_accepts_knowledge_graph(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        viz._normalize_graph = MagicMock(return_value=GRAPH_DICT)
+        viz._visualize_network_plotly = MagicMock(return_value=MagicMock())
+        communities = {"node_assignments": {"e1": 0, "e2": 1}, "num_communities": 2}
+        with patch(
+            "semantica.visualization.kg_visualizer.ColorPalette.get_community_colors",
+            return_value=["#ff0000", "#00ff00"],
+        ):
+            viz.visualize_communities(kg, communities=communities)
+        viz._normalize_graph.assert_called_once_with(kg)
+
+    def test_visualize_centrality_accepts_knowledge_graph(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        viz._normalize_graph = MagicMock(return_value=GRAPH_DICT)
+        viz._visualize_network_plotly = MagicMock(return_value=MagicMock())
+        viz.visualize_centrality(kg, centrality={"centrality": {}})
+        viz._normalize_graph.assert_called_once_with(kg)
+
+    def test_visualize_entity_types_accepts_knowledge_graph(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        viz._normalize_graph = MagicMock(return_value=GRAPH_DICT)
+        sys.modules["plotly.express"].bar.return_value = MagicMock()
+        viz.visualize_entity_types(kg)
+        viz._normalize_graph.assert_called_once_with(kg)
+
+    def test_visualize_relationship_matrix_accepts_knowledge_graph(self):
+        kg = self._make_kg()
+        viz = _make_viz()
+        viz._normalize_graph = MagicMock(return_value=GRAPH_DICT)
+        sys.modules["plotly.graph_objects"].Figure.return_value = MagicMock()
+        sys.modules["plotly.graph_objects"].Heatmap.return_value = MagicMock()
+        viz.visualize_relationship_matrix(kg)
+        viz._normalize_graph.assert_called_once_with(kg)
+
+    def test_knowledge_graph_importable_from_kg_module(self):
+        if self.KnowledgeGraph is None:
+            self.skipTest("semantica.kg.KnowledgeGraph not available")
+        try:
+            import semantica.kg as _kg_module
+            _ = _kg_module.KnowledgeGraph
+        except (ImportError, AttributeError) as exc:
+            self.fail(f"KnowledgeGraph not exported from semantica.kg: {exc}")
+
+    def test_knowledge_graph_empty_defaults(self):
+        kg = self.KnowledgeGraph()
+        self.assertEqual(kg.entities, [])
+        self.assertEqual(kg.relationships, [])
+        self.assertFalse(bool(kg))
+
+    def test_knowledge_graph_len(self):
+        kg = self._make_kg()
+        self.assertEqual(len(kg), len(ENTITIES))
+
+
 if __name__ == "__main__":
     unittest.main()