legacy-doc

legacy-doc is a small, dependency-free Python library for extracting readable text from classic Microsoft Word .doc files.

It is intentionally narrow:

• supports OLE Compound File Word 97-2003 .doc files
• extracts normalized text from the Word piece table
• rejects encrypted, malformed, non-OLE, or unsafe files clearly
• does not execute macros, embedded objects, links, scripts, or external references
• does not use native binaries, LibreOffice, antiword, or third-party parsers

Install

pip install legacy-doc

Usage

from legacy_doc import extract_text

with open("document.doc", "rb") as file:
    result = extract_text(file.read())

print(result.text)
print(result.metadata)

metadata always includes extracted text size fields and lightweight OLE inventory fields. When present in the document's saved Word metadata, it can also include values such as title, author, company, page count, word count, character count, creation time, and last-saved time.

API

extract_text(document_bytes: bytes, *, options: ExtractionOptions | None = None) -> DocExtractionResult

DocExtractionResult contains:

• text: normalized extracted text
• parser: parser name
• version: parser version
• metadata: lightweight extraction metadata
• warnings: non-fatal parser warnings

Example metadata:

{
    "chars": 1200,
    "bytes": 1225,
    "title": "Quarterly Notes",
    "author": "Liam Corrigan",
    "page_count": 4,
    "word_count": 210,
    "has_macros": False,
    "has_embedded_objects": False,
}

Limitations

This is not a full Microsoft Word renderer. It aims to recover readable text safely from common legacy .doc files. Formatting fidelity, OCR, images, macros, tracked changes, and embedded object extraction are out of scope.

page_count is read from saved Word document metadata when available. It is not recalculated by rendering the document, so it can be stale if the file was saved without updating document statistics.

Development

python -m pip install -e ".[dev]"
python -m pytest
python -m build
python -m twine check dist/*

The test suite includes small generated OLE WordDocument fixtures and malformed corpus cases for parser limits, encryption markers, missing piece tables, and cyclic FAT chains.

Security

legacy-doc treats input files as untrusted binary data. It performs bounded parsing, rejects encrypted documents, and never executes macros or embedded content. Report security issues privately before opening a public issue.