skandavivek / web-qa

ChatGPT Powered Question Answering Over Documents

AnswerChatAI has a unique feature that allows it to find “needles in information haystacks”. The app can extract relevant information from large volumes of text, even if the answer is buried deep within the document. Users can ask specific questions, and AnswerChatAI will locate the answer using OpenAI embeddings.

To run:

# create and activate python environment
$ python -m venv .venv
$ .\.venv\Scripts\activate

# change directory
$ cd open-ai-app/

# install project requirements
$ pip install -r requirements.txt

# start application
$ flask run

# NOTE: All app configurations are inside .env file

git stuff:

• git init
• git add .
• git commit -m "initial commit"
• git remote add origin https://github.com/skandavivek/web-qa.git
• git push -u origin main

Things to improve:

• urls not just starting from http (maybe best to separate URL from text)
• possible issues with non standard string characters in free response
• removing saving of files and instead directly accessing scraped file
• session cookies can't read large text (can do upto 847 words, so not bad...)