Catalogued footage with searchable metadata, auto-extracted thumbnails, and sortable indices. The library is the central data store that feeds chapter detection, dialogue matching, and rough cut assembly.
Source library — 20 entries with whisper_segments and chapter data. Raw clip metadata, durations, and transcription results.
Flat searchable index — 27 chapters with editorial_scores, search_text, and cross-references. Optimized for NLP queries.
AnaBot-generated topic index with Bruce Block taxonomy — shot_type, visual_component, story_structure, coverage_type.
Full-text search across all catalogued content. Queries run against segment text, chapter titles, and metadata fields. Results return with timestamps, scores, and clip references — ready for timeline insertion.
| Query Type | Example | Returns |
|---|---|---|
| Content search | "find all mentions of Scrooge" | Matching segments with timestamps |
| Chapter search | "chapters with score above 6" | Filtered chapter list with metadata |
| Clip search | "all clips from session 3" | Clip entries with full metadata |
| Taxonomy search | "close-up shots with high pacing" | Hashtag-indexed matches |
27 PNG thumbnails auto-extracted at chapter midpoints via FFmpeg. Used across the catalogue browser, chapter navigator, and library search interfaces.
| Parameter | Value |
|---|---|
| Count | 27 thumbnails |
| Format | PNG (lossless) |
| Extraction | FFmpeg midpoint seek per chapter |
| Source | A Christmas Carol read-through footage |
Auto-extracted frames from the catalogued footage library. Each clip is indexed, searchable, and scored.