How funstory-ai/babeldoc works
BabelDOC is a PDF scientific-paper translation library with CLI and Python entry points, producing translated and bilingual PDFs. The principal flow configures a translator and document-layout model, parses PDF content into an intermediate document representation, detects layout and paragraphs, translates text, then typesets and writes output. The README also documents page splitting, table translation, OCR-related handling, glossaries, and watermark options. Several internal links below are supported only by the source index or README; unsampled implementations and uncertain calls are not inferred.
Main entry points: high_level.py PDF translation workflow · translation_config.py Translation config · translator.py Translation engine · main.py CLI
Components 22
Entry and configuration
- CLI
babeldoc/main.py - Translation config
babeldoc/format/pdf/translation_config.py - Page splitting
babeldoc/format/pdf/split_manager.py - PDF translation workflow
babeldoc/format/pdf/high_level.py
PDF understanding
- Layout detection
babeldoc/docvision/base_doclayout.py - Layout RPC
babeldoc/docvision/rpc_doclayout6.py - PDF parsing
babeldoc/format/pdf/new_parser/native_parse.py - Character extraction
babeldoc/format/pdf/document_il/utils/extract_char.py - Document model
babeldoc/format/pdf/document_il/il_version_1.py - Paragraph detection
babeldoc/format/pdf/document_il/midend/paragraph_finder.py - Table detection
babeldoc/format/pdf/document_il/midend/table_parser.py - Scan detection
babeldoc/format/pdf/document_il/midend/detect_scanned_file.py
Translation
- Glossary
babeldoc/glossary.py - Translation engine
babeldoc/translator/translator.py - Paragraph translation
babeldoc/format/pdf/document_il/midend/il_translator.py - LLM translation
babeldoc/format/pdf/document_il/midend/il_translator_llm_only.py
PDF output
- Font mapping
babeldoc/format/pdf/document_il/utils/fontmap.py - Typesetting
babeldoc/format/pdf/document_il/midend/typesetting.py - PDF writer
babeldoc/format/pdf/document_il/backend/pdf_creater.py
Other components
- CLI or API user
- Translated PDF files
- Translation provider
Connections 25
- CLI or API user → CLI: runs
README.md - CLI → Translation config: configures
babeldoc/main.py - CLI → PDF translation workflow: starts
babeldoc/main.py - Translation config → Page splitting: selects strategy
babeldoc/format/pdf/translation_config.py - Translation config → Glossary: holds glossary
babeldoc/format/pdf/translation_config.py - Translation config → Translation engine: configures
babeldoc/format/pdf/translation_config.py - Translation config → Layout detection: configures
babeldoc/format/pdf/translation_config.py - PDF translation workflow → PDF parsing: parses PDF
babeldoc/format/pdf/high_level.py - PDF parsing → Character extraction: extracts text
babeldoc/format/pdf/document_il/utils/extract_char.py - Character extraction → Document model: populates
babeldoc/format/pdf/document_il/utils/extract_char.py - PDF translation workflow → Layout detection: detects layout
babeldoc/format/pdf/high_level.py - Layout detection → Layout RPC: delegates inference
babeldoc/docvision/base_doclayout.py - PDF translation workflow → Scan detection: checks scans
babeldoc/format/pdf/high_level.py - PDF translation workflow → Paragraph detection: finds paragraphs
babeldoc/format/pdf/high_level.py - PDF translation workflow → Table detection: parses tables
babeldoc/format/pdf/high_level.py - Paragraph detection → Document model: updates
babeldoc/format/pdf/document_il/midend/paragraph_finder.py - PDF translation workflow → Paragraph translation: translates content
babeldoc/format/pdf/high_level.py - PDF translation workflow → LLM translation: selects translation
babeldoc/format/pdf/high_level.py - Paragraph translation → Translation engine: requests translation
- LLM translation → Translation engine: calls LLM translation
babeldoc/format/pdf/document_il/midend/il_translator_llm_only.py - Translation engine → Translation provider: sends requests
babeldoc/translator/translator.py - PDF translation workflow → Font mapping: maps fonts
babeldoc/format/pdf/high_level.py - PDF translation workflow → Typesetting: typesets pages
babeldoc/format/pdf/high_level.py - PDF translation workflow → PDF writer: writes PDFs
babeldoc/format/pdf/high_level.py - PDF writer → Translated PDF files: produces