DocQA

Document Question Answering with the added powers of OCR. Also has it's own "Vector Database", no Pinecone needed.

Requirements

1. Install tesseract using brew -- brew install tesseract
2. Have tested on python 3.9 and 3.10, might break for other versions
3. OpenAI API keys with enought monies in it. Though 3.5 is honestly very cheap for personal small scale usage.
4. Install python libraries using requirements.txt

Usage

QA with a document

$ python main.py --filename [FILENAME] --api_key [OPENAI API KEY]

Changelog

29 May - Added chat session feature

Road ahead:

1. Chat session for docs - ✅
2. Multiple document inputs - TODO
3. Across sessions query - TODO