Audit findings & fixes:
- temperature was never set (API defaulted to 1.0) → now explicit 0.0 for deterministic JSON
- llm_max_tokens=65536 exceeded hard_limit=32768 → aligned to 32768
- Output ratio estimates were 5-30x too high (based on actual pipeline data):
stage2: 0.6→0.05, stage3: 2.0→0.3, stage4: 0.5→0.3, stage5: 2.5→0.8
- request_params now structured as api_params (what's sent to LLM) vs pipeline_config
(internal estimator settings) — no more ambiguous 'hard_limit' in request params
- temperature=0.0 sent on both primary and fallback endpoints
- _make_llm_callback now accepts request_params dict
- All 6 LLM call sites pass max_tokens, model_override, modality, response_model, hard_limit
- request_params stored in payload JSONB on every llm_call event (always, not just debug mode)
- Frontend JSON export includes full payload + request_params at top level
- DebugPayloadViewer shows 'Request Params' section even with debug mode off
- Answers whether max_tokens is actually being sent on pipeline requests
- Search now runs semantic + keyword in parallel, merges and deduplicates
- Keyword results always included with match_context explaining WHY matched
- Semantic results filtered by minimum score threshold (0.45)
- match_context shows 'Creator: X', 'Tag: Y', 'Title match', 'Content: ...'
- Qdrant points use deterministic uuid5 IDs (no more duplicates on reindex)
- Embedding timeout raised from 300ms to 2s (Ollama needs it)
- _enrich_qdrant_results reads creator_name from payload before DB fallback
- Frontend displays match_context as highlighted bar on search result cards
Two fixes:
1. page_moment_indices was referenced before assignment in the page
persist loop — moved assignment to top of loop body. This caused
"cannot access local variable" errors on every stage 5 run.
2. Stage 5 now catches LLMTruncationError and splits the chunk in
half for retry, instead of blindly retrying the same oversized
prompt. This handles categories where synthesis output exceeds
the model context window.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When the LLM splits a category group into multiple technique pages,
moments were blanket-linked to the last page in the loop, leaving all
other pages as orphans with 0 key moments (48 out of 204 pages affected).
Added moment_indices field to SynthesizedPage schema and synthesis prompt
so the LLM explicitly declares which input moments each page covers.
Stage 5 now uses these indices for targeted linking instead of the broken
blanket approach. Tags are also computed per-page from linked moments
only, fixing cross-contamination (e.g. "stereo imaging" tag appearing
on gain staging pages).
Deleted 48 orphan technique pages from the database.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Three resilience improvements to the pipeline:
1. LLMResponse(str) subclass carries finish_reason metadata from the LLM.
_safe_parse_llm_response detects truncation (finish=length) and raises
LLMTruncationError instead of wastefully retrying with a JSON nudge
that makes the prompt even longer.
2. Stage 4 classification now batches moments (20 per call) instead of
sending all moments in a single LLM call. Prevents context window
overflow for videos with many moments. Batch results are merged with
reindexed moment_index values.
3. run_pipeline auto-resumes from the last completed stage on error/retry
instead of always restarting from stage 2. Queries pipeline_events for
the most recent run to find completed stages. clean_reprocess trigger
still forces a full restart.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Stage 4 classification was truncating (finish=length) because the 0.15x
output ratio underestimated token needs. Inflated all stage ratios,
bumped the buffer from 20% to 50%, raised the floor from 2048 to 4096,
and fixed _safe_parse_llm_response to forward max_tokens on retry
instead of falling back to the 65k default.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Stage 3 (extraction) LLM calls now show the topic group label (e.g.,
'Sound Design Basics') and Stage 5 (synthesis) calls show the category
name. Displayed as a cyan italic label in the event row between the
event type badge and model name. Helps admins understand why there are
multiple LLM calls per stage.
Data model:
- New pipeline_runs table (id, video_id, run_number, trigger, status,
started_at, finished_at, error_stage, total_tokens)
- pipeline_events gains run_id FK (nullable for backward compat)
- Alembic migration 010_add_pipeline_runs
Backend:
- run_pipeline() creates a PipelineRun, threads run_id through all stages
- _emit_event() and _make_llm_callback() accept and store run_id
- Stage 6 (final) calls _finish_run() to mark complete with token totals
- mark_pipeline_error marks run as error
- Revoke marks running runs as cancelled
- Trigger endpoints pass trigger type (manual, clean_reprocess)
- New GET /admin/pipeline/runs/{video_id} — lists runs with event counts
- GET /admin/pipeline/events supports ?run_id= filter
Frontend:
- Expanded video detail now shows RunList instead of flat EventLog
- Each run is a collapsible card showing: run number, trigger type,
status badge, timestamps, token count, event count
- Latest run auto-expands, older runs collapsed
- Legacy events (pre-run-tracking) shown as separate collapsible section
- Run cards color-coded: cyan border for running, red for error,
gray for cancelled
- EventLog accepts optional runId prop to scope events to a single run
Deleting transcript_segments left the pipeline with nothing to process —
all stages would skip immediately. Segments come from the ingest step,
not from pipeline stages 2-6. Only pipeline_events and key_moments
(pipeline output) are deleted during clean reprocess.
- Backend: Video list now includes active_stage, active_stage_status, and
stage_started_at fields via DISTINCT ON subquery
- Backend: New GET /admin/pipeline/recent-activity endpoint returns
latest stage completions/errors with video context
- Frontend: 15-second auto-refresh with change detection — video rows
flash when status changes
- Frontend: Stage timeline dots on processing/complete/error videos
showing progress through stages 2-5, active stage pulses
- Frontend: Collapsible Recent Activity feed at top showing last 8
stage completions/errors with duration and creator
- Frontend: Bulk operation scrollable log showing per-video results
as they complete
- Frontend: Auto-refresh checkbox toggle in header
- Backend: POST /admin/pipeline/clean-retrigger/{video_id} endpoint that
deletes pipeline_events, key_moments, transcript_segments, and Qdrant
vectors before retriggering the pipeline
- Backend: QdrantManager.delete_by_video_id() for vector cleanup
- Frontend: Creator filter dropdown on pipeline admin page
- Frontend: Checkbox selection column with select-all
- Frontend: Bulk toolbar with Retrigger Selected and Clean Reprocess
actions, sequential dispatch with progress bar, cancel support
- Bulk dispatch uses 500ms delay between requests to avoid slamming API
773 key moments sat at 'pending' with 0 approved/edited/rejected.
review_status was never checked by any public-facing query — all content
was always visible regardless of review state.
Removed:
- backend/routers/review.py (10 endpoints)
- backend/tests/test_review.py
- frontend ReviewQueue, MomentDetail pages
- frontend client.ts (review-only API client)
- frontend ModeToggle, StatusBadge components
- Review link from AdminDropdown, Moments link from pipeline rows
- ReviewStatus, PageReviewStatus enums from models
- review_mode config flag
- review_status columns (migration 007)
- ~80 lines of mode-toggle CSS
Pipeline now always sets processing_status to 'published'.
Migration 007 drops columns, enums, and migrates 'reviewed' → 'published'.
Auto-mode commit 7aa33cd accidentally deleted 78 files (14,814 lines) during M005
execution. Subsequent commits rebuilt some frontend files but backend/, alembic/,
tests/, whisper/, docker configs, and prompts were never restored in this repo.
This commit restores the full project tree by syncing from ub01's working directory,
which has all M001-M007 features running in production containers.
Restored: backend/ (config, models, routers, database, redis, search_service, worker),
alembic/ (6 migrations), docker/ (Dockerfiles, nginx, compose), prompts/ (4 stages),
tests/, whisper/, README.md, .env.example, chrysopedia-spec.md
- 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)
