Lambda Document Interpreter Project

Current Status:

Project is set up and extraction + db store is working as expected. The main issue occured with sam local start-api runtime not being able to recognize the resulting output response claiming it's not a valid json. This is strange considering when running the project with uvicorn inside src/ using uvicorn main:app --reload --port=3000 it's working without issues so I am currently not sure of what's causing the issue in the lambda api side.

How to run:

1. Start the postgres database container

   docker compose up

2. Set up google api key in src/configs/settings.py in the GOOGLE_API_KEY field of the Settings class

3. Running using uvicorn:

   • Create a python virtual environment with python -m venv ./venv

   • Activate the virtual environment source venv/bin/activate

   • Install dependencies using pip install -r requirements.txt

   • Enter the src/ dir with cd src/

   • Run the project using uvicorn main:app --reload --port=3000

   • Send a request to the /extract route with something like (make sure to change "pdf_url" to an appropiate one):

     curl --request POST -H "Content-Type: application/json" -d '{"pdf_url": "http://192.168.122.254:8000/0809090-86.2024.8.12.0021.pdf", "case_id": "0809090-86.2024.8.12.0021"}' http://127.0.0.1:3000/extract

4. Running using sam (NOTE: this method for some reason didn't work with the current response format while the uvicorn runtime is working):

   • Make sure you're on the root path of this project and run:

     sam build && sam local start-api