- Added unmapped_type parameter to referenced_by_count sort
- This handles documents that don't have the field yet
- Updated ingest.py to include reference fields when indexing:
* internal_references
* internal_references_count
* referenced_by
* referenced_by_count
- Updated index mapping to include reference fields
- Documents without the field will sort as 0 (appear last)
Fixes BadRequestError: No mapping found for [referenced_by_count]
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Previously, secondary transcript parts were not being indexed
into Elasticsearch, causing the frontend to receive empty arrays
and display zero timestamps.
Changes:
- Add transcript_secondary_parts to Elasticsearch mapping
- Include secondary parts in bulk indexing actions
- Build secondary_full text from parts if not provided
- Match primary transcript structure (nested with start/duration/text)
Note: Existing data needs to be re-indexed for this fix to apply
to videos that already have secondary transcripts.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
