finatawa / aisummary

Simple Python code that lets you take long PDFs and generate summaries using GPT3.

Summarize PDFs Using GPT3

Original by Zed Tarar

Improved by Kai Turanski

Python code to help you skim through PDFs

This program allows you to select a PDF, divides it into chunks readable by GPT3, and creates a summary of it next to the original PDF.

Current limitations

• GPT3 may have a hard time interpreting paragraphs that bleed from one chunk to another
• Text output can sometimes contain badly formatted text, which can usually be deleted easily with no harm to comprehension

Editing

1. To edit the python file, make sure you have the right libraries installed. sys, math, and os should be local to your computer. Here are links for the openai, PyPDF2, and tkinter libraries.
2. You'll need to get your API key from OpenAI (ask chat GPT how to do that), then copy it into the script where indicated
3. The output is both printed on the command line and saved to summary of <your document name>.txt

Usage

1. Navigate to your ./dist/<your operating system>_version_pdf_extract/<your operating system>_version_pdf_extract.<extension> file
2. Double click to run the file and you're done!

Contribute

There are probably more elegant ways to create summaries from long PDFs, so if you have them, please share!

---

Original author LinkedIn

Improvement author LinkedIn