Chat With PDF

Simple application using Python and Streamlit using OpenAI API. Upload any pdf file and ask questions about it! It also supports images scans.

Screenshots

With History and Auto generated questions

Local environment

Make sure you have python and pip installed

1. Clone the repo

git clone https://github.com/pratikkalein/chat-with-pdf

2. Create a virtual environment

python3 -m venv venv

3. Activate the enviornment

source /venv/bin/activate

4. Install the requirments

pip install -r requirements.txt 

5. Create .env file at the root of the project and add your Open AI API key.

OPEN_AI_API_KEY=yourapikey

6. Run

streamlit run app.py

Deploy to Google Cloud

Build Docker Image using Cloud Build

gcloud builds submit --tag gcr.io/gcp-project-id/chat-with-pdf  --project=gcp-project-id

Deploy latest image to Cloud Run

Before running the command make sure you add your secret to the Secret Manager. Syntax for the --update-secrets

 --update-secretes=ENV_VAR_NAME=SECRET_NAME:VERSION

gcloud run deploy chat-with-pdf --image gcr.io/gcp-project-id/chat-with-pdf --platform managed  --project=gcp-project-id --allow-unauthenticated --region=asia-south1 --max-instances=2 --update-secrets=OPENAI_API_KEY=openai:1

If you wish to modify your deployment you can refere the documentation here.