Skip to content

README.md

Latest commit

 

History

History
138 lines (98 loc) · 5.29 KB

README.md

File metadata and controls

138 lines (98 loc) · 5.29 KB

office-oxide for Python — The Fastest Office Document Library for Python

The fastest Python library for text extraction, Markdown conversion, and editing across all six Microsoft Office formats. Powered by a Rust core via PyO3. Up to 100× faster than python-docx, openpyxl, and python-pptx. 100% pass rate on valid Office files — zero failures on legitimate Word/Excel/PowerPoint documents. MIT / Apache-2.0 licensed.

PyPI License: MIT OR Apache-2.0

Part of the office_oxide toolkit. Same Rust core, same pass rate as the Rust, Go, JavaScript (native), C# / .NET, and WASM bindings.

Quick Start

pip install office-oxide
from office_oxide import Document, EditableDocument, extract_text, to_markdown

# One-shot helpers
text = extract_text("report.docx")
md = to_markdown("spreadsheet.xlsx")

# Context-managed document
with Document.open("slides.pptx") as doc:
    print(doc.format)          # "pptx"
    print(doc.plain_text())
    print(doc.to_markdown())
    ir = doc.to_ir()           # format-agnostic IR as nested dicts

Why office-oxide?

  • Fast — 0.8ms mean DOCX, 5.0ms mean XLSX, 0.7ms mean PPTX; up to 100× faster than python-docx / openpyxl / python-pptx
  • Reliable — 100% pass rate on valid Office files (6,062-file corpus); zero failures on legitimate documents
  • Complete — 6 formats: DOCX, XLSX, PPTX + legacy DOC, XLS, PPT — one library
  • Permissive — MIT / Apache-2.0, unlike many alternatives that require GPL or AGPL
  • Drop-in — extract_text() and to_markdown() replace python-docx / openpyxl in one line

Performance

Benchmarked on 6,062 files from 11 independent public test suites. Single-thread, release build, warm disk cache.

Library Format Mean Pass Rate License
office_oxide DOCX 0.8ms 98.9% MIT
python-docx DOCX 11.8ms 95.1% MIT
office_oxide XLSX 5.0ms 97.8% MIT
python-calamine XLSX 13.9ms 96.6% MIT
openpyxl XLSX 94.5ms 96.2% MIT
office_oxide PPTX 0.7ms 98.4% MIT
python-pptx PPTX 32.5ms 86.7% MIT

Full methodology and corpus breakdown in BENCHMARKS.md.

Installation

pip install office-oxide

Pre-built wheels for Linux (x86_64, aarch64, musl), macOS (x86_64, arm64), and Windows (x86_64). Python 3.8–3.14. No system dependencies, no Rust toolchain required.

API

Extraction

from office_oxide import Document, extract_text, to_markdown, to_html

# One-shot helpers
text = extract_text("report.docx")
md   = to_markdown("spreadsheet.xlsx")
html = to_html("slides.pptx")

# Document object
doc = Document.open("presentation.pptx")
text = doc.plain_text()
md   = doc.to_markdown()
html = doc.to_html()
ir   = doc.to_ir()           # structured dict (format-agnostic IR)
fmt  = doc.format            # "pptx"

Document.open accepts both str and pathlib.Path. Can also be used as a context manager.

From bytes

with open("file.docx", "rb") as f:
    doc = Document.from_bytes(f.read(), "docx")

Editing

with EditableDocument.open("template.docx") as ed:
    ed.replace_text("{{NAME}}", "Alice")
    ed.save("out.docx")

Spreadsheet cells

with EditableDocument.open("data.xlsx") as ed:
    ed.set_cell(0, "A1", "Revenue")
    ed.set_cell(0, "B1", 12345.67)
    ed.set_cell(0, "C1", True)
    ed.save("data.edited.xlsx")

Other languages

office_oxide ships the same Rust core through six bindings:

Why I built this

I needed a library that could read all six Office formats at once — not six separate packages — and I needed it without pulling in a JVM or a GPL-licensed dependency. Nothing existed that combined speed, correctness, and a permissive license across the full DOCX / XLSX / PPTX / DOC / XLS / PPT surface, so I wrote it in Rust and wrapped it for every language I use day-to-day.

If it saves you a dependency or a license audit, a star on GitHub genuinely helps. If something's broken or missing, open an issue.

— Yury

License

Dual-licensed under MIT or Apache-2.0 at your option.

Python + Rust core | MIT / Apache-2.0 | 100% pass rate on valid Office files (6,062-file corpus) | Up to 100× faster than alternatives | 6 formats