Extracting PDF Tables with Camelot: A Python Tutorial

PDF files are widely used for formal documents, but extracting tabular data from them can be painful, especially when the tables are embedded as text.

Camelot is a Python tool that extracts tables from PDF files and returns them as pandas DataFrames, allowing users to work with the data just like they would with CSV files.

Camelot Overview

The project’s GitHub address is https://github.com/camelot-dev/camelot . Users can import the library, read a PDF, and export the extracted tables to various formats such as CSV, JSON, Excel, HTML, or SQLite.

Code Example

<code>&gt;&gt;&gt; import camelot<br>&gt;&gt;&gt; tables = camelot.read_pdf('foo.pdf')  # similar to pandas.read_csv<br>&gt;&gt;&gt; tables[0].df  # get a pandas DataFrame!<br>&gt;&gt;&gt; tables.export('foo.csv', f='csv', compress=True)  # export to CSV (or json, excel, html, sqlite)<br>&gt;&gt;&gt; tables[0].to_csv('foo.csv')  # also supports to_json, to_excel, to_html, to_sqlite<br>&gt;&gt;&gt; tables<br>&lt;TableList n=1&gt;<br>&gt;&gt;&gt; tables[0]<br>&lt;Table shape=(7, 7)&gt;  # shows the extracted table dimensions<br>&gt;&gt;&gt; tables[0].parsing_report<br>{'accuracy': 99.02, 'whitespace': 12.24, 'order': 1, 'page': 1}</code>

The output handles merged cells by inserting empty rows, providing a reliable way to preserve table structure.

Installation Methods

The author provides three ways to install Camelot:

Using conda (the simplest): conda install -c conda-forge camelot-py

Using pip (most popular): pip install camelot-py[cv]

Cloning the repository and installing from source: git clone https://www.github.com/camelot-dev/camelot<br>cd camelot<br>pip install ".[cv]"

These methods allow users to quickly set up Camelot and start extracting tables from PDFs for data analysis or further processing.