pdfplumber-rs

High-performance PDF text, table, and geometry extraction for Python — powered by Rust.

pdfplumber-rs is a Rust-native reimplementation of pdfplumber exposed to Python via PyO3. The Rust extension is installed as the private pdfplumber._native submodule so the public package can provide a Python compatibility layer separately.

Installation

pip install pdfplumber-rs

Quick Start

import pdfplumber

# Open a PDF
pdf = pdfplumber.open("document.pdf")

# Access pages
for page in pdf.pages:
    # Extract text
    text = page.extract_text()
    print(text)

    # Extract words with bounding boxes
    words = page.extract_words()
    for w in words:
        print(f"{w['text']} at ({w['x0']}, {w['top']}, {w['x1']}, {w['bottom']})")

    # Extract tables
    tables = page.find_tables()
    for table in tables:
        rows = table.extract()
        for row in rows:
            print(row)

# Open from bytes
from pdfplumber import _native

with open("document.pdf", "rb") as f:
    pdf = _native.PDF.open_bytes(f.read())

The top-level pdfplumber.open alias is available for filesystem paths. Other compatibility-facade work remains in progress; use the private extension explicitly for the native-only API shown above.

Passing laparams={} to pdfplumber.open enables native horizontal layout analysis. The resulting textboxhorizontal and textlinehorizontal lists participate in the same page/document object caches and serialization as ordinary objects, and are also available through .textboxhorizontals and .textlinehorizontals. Horizontal grouping honors line_overlap, char_margin, word_margin, and line_margin; vertical layout and embedded-figure hierarchy support remain in progress.

API Reference

PDF

Page

Table

CroppedPage

Supports the same content surface as Page: chars, lines, rects, curves, images, extract_text(), extract_words(), find_tables(), extract_tables(), and point2coord(), plus further crop(), within_bbox(), and outside_bbox().

Rust-Native Extensions

Python pdfplumber v0.11.10 does not define high-level document APIs for bookmarks, forms, signatures, structural validation, or image-byte extraction. pdfplumber-rs exposes these only through document.rust, so they cannot be mistaken for compatibility behavior or silently collide with future upstream methods. Page indexes and bookmark destinations in this namespace retain the Rust API's 0-based convention.

Comparison with Python pdfplumber

Performance

pdfplumber-rs benefits from Rust's zero-cost abstractions and compiled performance:

• Text extraction: Typically 5-20x faster than Python pdfplumber
• Table detection: Typically 3-10x faster for lattice-based tables
• Memory usage: Lower memory footprint due to Rust's ownership model

Actual speedups depend on document complexity and system configuration.

License

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