Feat/dashboard UI redesign (#65)

* fix(mypy/ruff): correct extractor extend path and sync dev dependencies

- Fix extractor/pyproject.toml: change `../../pyproject.toml` to `../pyproject.toml`
  so ruff/mypy/coverage/pytest tool configs correctly extend from the root
- Run `uv sync --all-extras` to install httpx and prometheus-client, resolving
  mypy `import-not-found` errors for dashboard service

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

* feat(dashboard): redesign UI with dark Tailwind theme from Stitch design

- Complete rewrite of index.html: dark theme via Tailwind CDN, Inter/JetBrains
  Mono fonts, Material Symbols icons; logo in #app-logo div with prominent
  comment block for easy swapping
- Complete rewrite of dashboard.js: remove Chart.js dependency, replace with
  CSS height bars (queue size metrics) and SVG stroke-dashoffset circular gauges
  (processing rates); WebSocket connection and all real-time data updates preserved
- Simplify styles.css: retain only base reset and .connection-status /
  .status-indicator / .log-entry stubs for Playwright test selector compatibility
- Update test_dashboard_ui.py: align selectors with new HTML structure
  (#service-{name} IDs, #db-neo4j, #db-postgresql, Queue Size Metrics / Processing
  Rates section headings, responsive design viewport checks)

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

* chore(extractor): remove pyproject.toml from pure Rust project

The extractor is a Rust binary with no Python code, no Python tests, and is
intentionally excluded from the uv workspace and coverage source tracking.
Having a pyproject.toml was misleading and caused ruff to fail when the
extend path resolved to a non-existent parent directory.

- Delete extractor/pyproject.toml
- scripts/update-project.sh: remove extractor/*/pyproject.toml from
  suggested git add command
- scripts/README.md: fix component counts (8→7 Python files, 15→14 total),
  remove stale extractor/pyextractor and extractor/extractor directory
  references, list actual Python service directories

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

* docs: update all docs to reflect current architecture

Remove all references to the removed Discovery/ML service and the
former Python extractor across all documentation files.

README.md
- Replace "AI Discovery Engine" bullet with "Interactive Explorer"
- Remove "ML-Powered Discovery" from Key Features

dashboard/README.md
- Remove discovery service from monitored services list and WebSocket
  JSON example
- Remove DISCOVERY_URL env var
- Fix POSTGRES_ADDRESS: postgres:5433 → postgres:5432 (internal Docker port)
- Update frontend stack: Tailwind CSS replacing "Vanilla JavaScript"
- Fix static file name: script.js → dashboard.js
- Remove false "WebSocket connections are authenticated" security claim

extractor/README.md
- Rewrite intro: remove "reimplementation of the Python-based extractor"
  framing — this IS the extractor
- Remove "Integration" section's Python extractor references
- Remove stale "Migration from Python Extractor" section entirely
- Fix coverage command: cargo-tarpaulin → cargo-llvm-cov

docs/monitoring.md
- Fix "Python Extractor" → "Extractor" in log example
- Replace all `uv run task X` with `just X` (taskipy is not installed;
  project uses justfile)
- Update Last Updated date to 2026-02-21

docs/troubleshooting.md
- Rename "Python/Extractor Download Failures" → "Extractor Download Failures"
- Remove ML model cache chown block from Permission Denied section
- Fix health check list: remove Discovery ports 8004/8005, add Explore 8007
- Fix port conflict netstat/lsof: replace 8004/8005 with 8006/8007
- Remove "ML model operations" from DEBUG level log categories
- Fix broken link: logging-configuration.md → logging-guide.md
- Rewrite "Neo4j Schema Warnings" to not mention Discovery service
- Remove entire "Discovery Service Issues" section (asyncpg, cache dirs,
  TRANSFORMERS_CACHE — all artifacts of a removed ML service)
- Rename "Extractor Issues (Python/Rust)" → "Extractor Issues"
- Remove "Switch to Extractor (20-400x faster)" note — Rust IS the extractor
- Fix "Discovery service timeouts" → "Explore service timeouts"

docs/recent-improvements.md
- Fix batch processing config defaults: 100/5.0 → 500/2.0 (actual deployed values)
- Add Dashboard UI Redesign entry documenting the Tailwind/SVG/CSS-bar redesign

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

---------

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: SimplicityGuy

README.md

@@ -32,7 +32,7 @@ Discogsography transforms monthly Discogs data dumps (50GB+ compressed XML) into
 
 - **🔗 Neo4j Graph Database**: Navigate complex music industry relationships
 - **🐘 PostgreSQL Database**: High-performance queries and full-text search
-- **🤖 AI Discovery Engine**: Intelligent recommendations and analytics
+- **🔍 Interactive Explorer**: Graph visualisation, trends, and path discovery
 - **📊 Real-time Dashboard**: Monitor system health and processing metrics
 
 Perfect for music researchers, data scientists, developers, and music enthusiasts who want to explore the world's largest music database.
@@ -109,7 +109,6 @@ graph TD
 - **🐋 Container Security**: Non-root users, read-only filesystems, dropped capabilities
 - **📝 Type Safety**: Full type hints with strict mypy validation and Bandit security scanning
 - **✅ Comprehensive Testing**: Unit, integration, and E2E tests with Playwright
-- **🧠 ML-Powered Discovery**: Semantic search, graph algorithms (PageRank, community detection)
 
 ## 🚀 Quick Start
 

dashboard/README.md

@@ -6,7 +6,7 @@ Real-time monitoring dashboard for the Discogsography system with WebSocket supp
 
 The dashboard service provides a web-based interface to monitor all Discogsography services, including:
 
-- Service health status (extractor, graphinator, tableinator, discovery)
+- Service health status (extractor, graphinator, tableinator)
 - RabbitMQ queue metrics and consumer counts
 - PostgreSQL and Neo4j database statistics
 - Real-time activity logs
@@ -15,7 +15,7 @@ The dashboard service provides a web-based interface to monitor all Discogsograp
 ## Architecture
 
 - **Backend**: FastAPI with WebSocket support
-- **Frontend**: Vanilla JavaScript with dynamic updates
+- **Frontend**: Tailwind CSS, Inter/JetBrains Mono fonts, SVG circular gauges, CSS bar charts
 - **Port**: 8003 (configurable via `DASHBOARD_PORT`)
 - **Health Endpoint**: `/health` (port 8003)
 
@@ -32,14 +32,13 @@ DASHBOARD_PORT=8003       # Port for web interface
 EXTRACTOR_URL=http://extractor:8000
 GRAPHINATOR_URL=http://graphinator:8001
 TABLEINATOR_URL=http://tableinator:8002
-DISCOVERY_URL=http://discovery:8004
 
 # Database connections
 NEO4J_ADDRESS=neo4j:7687
 NEO4J_USERNAME=neo4j
 NEO4J_PASSWORD=discogsography
 
-POSTGRES_ADDRESS=postgres:5433
+POSTGRES_ADDRESS=postgres:5432
 POSTGRES_USERNAME=discogsography
 POSTGRES_PASSWORD=discogsography
 POSTGRES_DATABASE=discogsography
@@ -67,8 +66,7 @@ The dashboard broadcasts updates every 5 seconds with the following data:
     "services": {
       "extractor": { "status": "healthy", "health_url": "..." },
       "graphinator": { "status": "healthy", "health_url": "..." },
-      "tableinator": { "status": "healthy", "health_url": "..." },
-      "discovery": { "status": "healthy", "health_url": "..." }
+      "tableinator": { "status": "healthy", "health_url": "..." }
     },
     "queues": {
       "labels": { "messages": 0, "consumers": 1 },
@@ -116,9 +114,9 @@ uv run pytest tests/dashboard/test_dashboard_api_integration.py -v
 
 The frontend consists of static files in the `static/` directory:
 
-- `index.html` - Main dashboard page
-- `styles.css` - Dashboard styling
-- `script.js` - WebSocket client and UI updates
+- `index.html` - Main dashboard page (Tailwind CSS, dark theme)
+- `styles.css` - Base reset and legacy selector stubs
+- `dashboard.js` - WebSocket client and UI update logic
 
 ## Docker
 
@@ -139,6 +137,5 @@ The dashboard itself exposes metrics at `/api/metrics` for monitoring its own he
 ## Security
 
 - CORS is configured for production use
-- WebSocket connections are authenticated
 - All external service calls have timeouts
 - Sensitive configuration is loaded from environment variables