gsacavdm / python-pdf

Python scripts to parse PDF documents

Overview

This program is meant for extraction of data from many PDFs which have the same structure, specifically, that their lines numbers won't change.

Examples - monthly statements, paystubs, insurance claims, etc.

How to Run

1. Copy the a sample PDF to the PDFs directory
2. Run generate_parsing_json.py as follows:

   ./generate_parsing_json -p pdfs/my-statement.pdf > parsing_jsons/statements.json
   

3. Manually edit the generated statements.json file as follows:
   • Remove any entries that you don't care about parsing or asserting
   • Keep intact those that you want to assert. Asserting is used as a safety check to ensure things that the structure of the document (i.e. the line numbers where things are at) hasn't changed.
   • Add a meaningful name that's a valid python variable name for each of the entries you want to extract.
   • Remove the assert_value for each of the entries you want to extract.
4. Run extract_pdf.py as follows:

   ./extract_pdf.py -p pdfs/my-statement.pdf -j parsing_jsons/statements.json > outputs/my-statement.json
   

5. You can then write a bash script to do this for all files in the pdfs folder. I should later modifh extract_pdf to support input and output directories as parameters.