admin
950284e236
This commit significantly improves the robustness and determinism of the Meeting Lab processing pipeline and establishes the first Release Candidate baseline for end-to-end evaluation. Highlights - BUG-009 - Implement deterministic responsible-party validation - Normalize participant aliases using Meeting Context - Reject invalid responsible values (dates, locations, technical terms, projects, products, unknown entities) - Record structured responsibility validation metadata - Add focused regression tests - BUG-010 - Implement adaptive num_predict estimation for Semantic Consolidator - Eliminate JSON truncation caused by fixed output limits - Add deterministic source coverage repair - Preserve strict post-repair validation - Add regression tests - BUG-011 - Implement Working Protocol V2 renderer contract enforcement - Preserve raw renderer responses - Reject invalid protocol output instead of accepting malformed documents - Add deterministic cleanup for harmless formatting deviations - Add focused renderer regression tests - Meeting Context - Validate Meeting Context V1 - Integrate authoritative participant alias normalization - Documentation - Update architecture documentation - Update output documentation - Update regression bug tracker The pipeline now fails safely instead of silently accepting invalid intermediate or final artifacts. Remaining work focuses primarily on extraction quality and semantic classification (decisions, action items, protocol faithfulness), rather than pipeline robustness.
Progeo Meeting
Real-Life Reference Meeting #2 for Meeting Lab.
This sample provides an additional real meeting dataset for evaluating the pipeline beyond the project-process meeting. It is intended for quality assessment, regression analysis and comparison of pipeline behaviour across different meeting structures.
Files
progeo_meeting_speech.json: original Whisper JSON transcript.progeo_meeting_speech_cleaned.json: cleaned Whisper JSON transcript generated withscripts/clean_whisper_json.py.meeting_context.yaml: Meeting Context V1 scaffold.
Meeting Context Status
The Meeting Context is intentionally incomplete. It contains only known meeting-level metadata and no inferred participants.
Participant, mentioned-person, organization, department, alias and known-entity metadata require manual completion before context-aware extraction should be treated as authoritative for this sample.
Known Characteristics
- Strongly uneven speaking shares.
- Comparatively stable discussion topics.
- Intended as an additional evaluation dataset.
Preparation Notes
The cleaned transcript was created with:
.venv/bin/python scripts/clean_whisper_json.py \
samples/real_live/progeo_meeting/progeo_meeting_speech.json \
-o samples/real_live/progeo_meeting/progeo_meeting_speech_cleaned.json
Cleaning result:
- Segments before: 1448
- Segments after: 1448
- Segments removed: 0