#EPUB
3 related articles
EPUB Conversion Pitfalls: Chapter Order Can't Be Guessed—The Spine Is the Only Truth
Developer shares EPUB-to-Markdown pitfalls: chapter order must come from the spine, not alphabetical sorting or manifest; footnotes need inlining; EPUB2 and EPUB3 nav files differ and need a shim. Final output is structured Markdown for RAG pipelines.
Parsing and Rebuilding EPUB with Python: Practical Lessons from an AI Book Translation Project
Based on the LectuLibre AI book translation project, this article details parsing and rebuilding EPUB files with Python's ebooklib and lxml: namespace pitfalls, text extraction and replacement strategies, batch translation workflows, encoding and corrupted file handling, and the unsolved challenge of preserving inline formatting.
EPUB Conversion Pitfall: Reading Order Lives in the .opf Spine, Not the Filenames
Why EPUB-to-Markdown converters scramble chapters: reading order is defined in content.opf's spine, not filenames. Learn the correct 3-step parsing flow and XML fault tolerance tips.