API reference

Signatures and member documentation on these pages are generated from the code (sphinx.ext.autodoc). Everything listed here is importable from doc2mark, except where a page names another module (doc2mark.ocr, doc2mark.ocr.schema, doc2mark.judge). The user guide explains how the pieces fit together; start with Quickstart.

• UnifiedDocumentLoader
  • UnifiedDocumentLoader
• Convenience functions
  • load()
  • document_to_markdown()
  • batch_convert_to_markdown()
  • batch_process_documents()
  • batch_process_files()
• Results, enums and exceptions
  • ProcessedDocument
  • DocumentMetadata
  • Enums
  • Exceptions
• OCR
  • The facade
  • Configuration
  • Results and enums
  • Providers
• Structured OCR Schema
  • The verbatim / interpretation boundary
  • OCRPage — the top-level result
  • RawExtraction — verbatim page content
  • Table — a transcribed table
  • KeyValue — a label/value pair
  • Metric — a typed numeric assertion
  • The nested interpretation models
  • Interpretation — the model’s reading
  • A complete filled OCRPage (receipt)
  • Helpers
• OCR cache
  • create_ocr_cache()
  • OCRCache
  • MemoryOCRCache
  • RedisOCRCache
  • NoOpOCRCache
  • CachedOCR
• Chunking
  • chunk_content()
  • ChunkingConfig
  • Chunk
• Judge
  • TypeSafeJudge
  • resolve_judge()
  • judge_hooks()