GitHub - camelot-dev/camelot: A Python library to extract tabular data from PDFs (original) (raw)

Camelot: PDF Table Extraction for Humans

Camelot is a Python library that can help you extract tables from PDFs.

Features

• 📊 Five parsers — lattice (ruled tables), stream (whitespace), the text-alignment network / hybrid, and the optional neural ml (Table Transformer) for hard borderless tables — plus flavor="auto" to pick one for you.
• 🤖 Borderless & scanned — the optional ml backend (pip install "camelot-py[ml]") recovers structure that the heuristic parsers can't on borderless tables; add [ocr] to read scanned / image-only PDFs with no text layer.
• 🧠 Vector + raster line detection — engine="combined" unions the PDF's native vector ruled lines with OpenCV detection, so faintly-ruled tables are still found.
• 🐼 pandas output — every table is a DataFrame, ready for analysis.
• 📤 Many export formats — CSV, JSON, Excel, HTML, Markdown, and SQLite.
• 📐 Quality metrics — accuracy, whitespace, and a confidence score per table; drop noise with TableList.filter(...).
• 🧩 Multi-page tables — stitch continuations across pages with stack_contiguous().
• 🎛️ Highly configurable — table areas/regions, column separators, text processing, and more.
• 🔌 Flexible input — a file path, URL, raw bytes, or any binary file-like object.
• 🖥️ CLI included — camelot lattice file.pdf, etc.
• 📦 Light install — the default pdfium backend is bundled, with no system dependencies.

Extract tables from PDFs in just a few lines of code:

Try it yourself in our interactive quickstart notebook.

Or check out a simple example using this pdf.

import camelot tables = camelot.read_pdf('foo.pdf') tables tables.export('foo.csv', f='csv', compress=True) # json, excel, html, markdown, sqlite tables[0] <Table shape=(7, 7)> tables[0].parsing_report { 'accuracy': 99.02, 'whitespace': 12.24, 'order': 1, 'page': 1 } tables[0].to_csv('foo.csv') # to_json, to_excel, to_html, to_markdown, to_sqlite tables[0].df # get a pandas DataFrame!

Camelot also comes packaged with a command-line interface!

Refer to the QuickStart Guide to quickly get started with Camelot, extract tables from PDFs and explore some basic options.

Tip: Visit the parser-comparison-notebook to get an overview of all the packed parsers and their features.

Note: The built-in parsers need a text-based PDF (as Tabula explains, "If you can click and drag to select text in your table in a PDF viewer, then your PDF is text-based"). For scanned / image-only PDFs, install the neural backend with OCR — pip install "camelot-py[ml,ocr]" — and use camelot.read_pdf(..., flavor="ml"): the model reads the structure from the page image and OCR supplies the text.

You can check out some frequently asked questions here.

Which parser should I use?

The ml backend keeps Camelot honest: the model only supplies the tablestructure, while cell text comes from the PDF's own text layer (or OCR for scans) — so it never invents or alters a value.

Why Camelot?

• Configurability: Camelot gives you control over the table extraction process with tweakable settings.
• Metrics: You can discard bad tables based on metrics like accuracy and whitespace, without having to manually look at each table.
• Output: Each table is extracted into a pandas DataFrame, which seamlessly integrates into ETL and data analysis workflows. You can also export tables to multiple formats, which include CSV, JSON, Excel, HTML, Markdown, and Sqlite.

See comparison with similar libraries and tools.

Installation

Camelot's default image-conversion backend is pdfium, which ships as a wheel — so a plain install needs no system dependencies. The optional ghostscript and poppler backends require additional dependencies.

Using uv

uv is a fast Python package and project manager. To add Camelot to a project:

Or to install it into the current environment:

uv pip install camelot-py

Using pip

Using conda

conda is the package manager for the Anaconda distribution:

conda install -c conda-forge camelot-py

From the source code

git clone https://github.com/camelot-dev/camelot.git cd camelot uv pip install "." # or: pip install "."

Optional extras

pip install "camelot-py[ml]" # neural flavor='ml' (Table Transformer; pulls PyTorch) pip install "camelot-py[ocr]" # OCR text source for scanned PDFs (use with [ml]) pip install "camelot-py[ml,ocr]" # both — borderless + scanned pip install "camelot-py[plot]" # matplotlib debug plots

The core install stays light: [ml]/[ocr] are imported lazily, so a plainimport camelot never loads PyTorch or OCR.

Documentation

The documentation is available at http://camelot-py.readthedocs.io/.

Contributing

The Contributor's Guide has detailed information about contributing issues, documentation, code, and tests.

Versioning

Camelot uses Semantic Versioning. For the available versions, see the tags on this repository. For the changelog, you can check out the releases page.

License

This project is licensed under the MIT License, see the LICENSE file for details.

The documentation theme is licensed under a seperate BSD-like License, see the LICENSE file for details.