Replaces outdated README with streamlined version covering all 8 services,
complete API endpoints, 6-stage information flow diagram, and current
project structure.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Use Docker embedded DNS (127.0.0.11) with 30s TTL and variable-based
proxy_pass so nginx re-resolves the API container IP after recreates
instead of caching the startup IP forever.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add content_hash (SHA-256 of transcript text) to source_videos (migration 005)
- 3-tier duplicate detection at ingest: exact filename, content hash,
then normalized filename + duration (handles yt-dlp re-downloads)
- Snapshot prior technique_page_ids to Redis before pipeline dispatch
- Stage 5 matches prior pages by creator+category before slug fallback,
enabling version snapshots on reprocessing even when LLM generates
different slugs
- Expose content_hash in API responses and admin pipeline dashboard
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add pipeline_events table (migration 004) for structured stage logging
- Add PipelineEvent model with token usage tracking
- Admin pipeline dashboard with video list, event log, worker status,
trigger/revoke controls, and collapsible JSON payload viewer
- Version switcher on technique pages — view historical snapshots
with pipeline metadata (model names, prompt hashes)
- Frontend types for pipeline admin and version APIs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Version dropdown appears when version_count > 0 (hidden until first re-run)
- Selecting a historical version overlays snapshot content (title, summary, body, chains, plugins)
- Key moments and related links always show live data (not versioned)
- Pipeline metadata block shows model, capture time, and prompt file hashes (truncated)
- Cyan banner when viewing historical version with "Back to current" button
- fetchTechniqueVersion API function for single version detail
- ContentReport model with generic content_type/content_id (supports any entity)
- Alembic migration 003: content_reports table with status + content indexes
- POST /reports (public), GET/PATCH /admin/reports (admin triage)
- Report modal on technique pages with issue type dropdown + description
- Admin reports page with status filter, expand/collapse detail, triage actions
- All CSS uses var(--*) tokens, dark theme consistent
- Creators: response_model changed from list to {items, total, offset, limit} matching frontend CreatorBrowseResponse
- Review queue: limit raised from 100 to 1000
- New GET /review/moments/{moment_id} endpoint for direct moment fetch
- MomentDetail uses fetchMoment instead of fetching full queue
- Merge candidates fetch uses limit=100
- max_tokens bumped from 16384 to 32768 (extraction responses still hitting limits)
- All GSD planning/completion artifacts for M002 (deployment) and M003 (DNS + LLM routing)
- KNOWLEDGE.md updated with XPLTD domain setup flow and container healthcheck patterns
- DECISIONS.md updated with D015 (subnet) and D016 (Ollama for embeddings)
- Changed network subnet from 172.24.0.0/24 (used by xpltd_docs) to 172.32.0.0/24
- Added chrysopedia-qdrant (qdrant/qdrant:v1.13.2) with healthcheck
- Added chrysopedia-ollama for embedding model (nomic-embed-text)
- Web UI exposed on 0.0.0.0:8096 (XPLTD port-suffix convention)
- API internal-only (via nginx proxy)
- Added HEALTHCHECK to Dockerfile.api
- Updated .env.example with FYN DGX LLM endpoint
