feat: Implement speaker analysis and EPUB 3 export functionality

- Added speaker analysis module to infer speaker identities from text chunks.
- Introduced SpeakerGuess and SpeakerAnalysis data classes for managing speaker data.
- Developed functions for analyzing speaker occurrences and confidence levels.
- Created EPUB 3 exporter to generate EPUB packages with synchronized narration and media overlays.
- Implemented configurable chunking options for TTS synthesis and EPUB alignment.
- Enhanced JavaScript for speaker preview functionality in the web interface.
- Added comprehensive tests for chunking and EPUB exporting features.
- Documented upgrade plan for transitioning to EPUB 3 with multi-speaker support.
This commit is contained in:
JB
2025-10-07 17:57:53 -07:00
parent bacf1b2f9e
commit 41f56a8491
18 changed files with 2844 additions and 14 deletions
+4
View File
@@ -1,3 +1,7 @@
# Unreleased
- Added an EPUB 3 packaging pipeline that builds media-overlay EPUBs from generated audio and chunk metadata.
- Persisted chunk timing metadata in job artifacts and exercised the exporter with automated tests.
# 1.1.9
- Fixed the issue where spaces were deleted before punctuation marks while generating subtitles.
- Fixed markdown TOC generation breaks when "Replace single newlines" is enabled.