Add X-Timezone, X-Language, X-Version headers to use new dataset serv…

…ice (#72)

Point Vite dev proxy at new datasets.app.trustgraph.ai service.
Fetch version from /config/version.json (gracefully skipped if unavailable)
and include timezone, language, and version headers on all demo-data fetches.

Add documents submitted section to Welcome page, show names and descr…

…iptions in workspace/collection/flow selectors (#71)

Add catalog/dataset section headings, search workflow card, remove Se…

…arch tab, show Admin tab (#70)

Show catalog thumbnail and landing page link on Welcome page (#69)

Import catalog triples into named graph urn:graph:catalog (#68)

Add dynamic navigation system, Welcome page with DCAT catalogs, Searc…

…h tab, and docs viewer (#67)

- Replace monolithic components.json with split config files (navigation.json,
  tabs.json, components.json, routes.json) and intent-based routing
- All routing is now config-driven through an intent router — no hardcoded routes
- Welcome page queries the knowledge graph for DCAT catalogs, showing datasets
  with thumbnails, keywords, access service buttons, and documentation links
- Search tab with embedding-based semantic search, type chips, keyword pills,
  thumbnails, and RawNodeDetailPanel side viewer for browsing the graph
- DocsViewer component renders markdown documentation from dcat:landingPage URLs
- Rename component IDs to be more descriptive (e.g. raw-graph → graph-navigator)
- Add thumbnail support to RawNodeDetailPanel via schema:image
- Demo data loader now handles the manifest "catalog" field for DCAT metadata

Add guidance help icon to demo data loader (#66)

Improve demo data loader: use marked for markdown rendering, fix imag…

…e sizing, rename plugin route to avoid proxy collision, update demo tab descriptions (#65)

Add demo data loader plugin with streaming Turtle parser and bulk imp…

…ort (#64)

- New plugin-demo-data package: dataset browser with thumbnail grid,
  hero detail view with markdown descriptions, and progress overlay
  for loading/test-parse operations
- Streaming Turtle parser using N3.js StreamParser fed from fetch
  response body, yielding quads via async generator
- SDL pipeline for structured data: multi-stage generator pipeline
  (parse CSV → preprocess → map/transform → format ExtractedRows)
- BulkApi updated to accept AsyncIterable for end-to-end streaming
  from fetch → parse → batch → WebSocket
- Workspace auto-refresh after creation via useWorkspaces().refetch
- Vite dev proxy for /demo-data to GitHub raw datasets
- Production --proxy CLI flag on trustgraph-ui for reverse proxying

Allow hidden tabs to be routed via direct URL (#63)

Hidden tabs are excluded from the nav bar but still accessible
by navigating directly to their URL path (e.g. /admin).