LC-Vision

This is an ongoing project being built alongside the LC Pill Checker project to expand its capabilities to allow for image analysis

Architecture

Environment Variables

To run this project, you will need to add the following environment variables to your .env file

REACT_APP_IMG_ENDPOINT- API endpoint to upload images to

REACT_APP_KEY_ENDPOINT - API endpoint to store the user's OpenAI API key

SECRET_KEY - for flask sessions configuration

REDIS_HOST - Redis cloud endpoint

REDIS_PWD - Redis cloud endpoint password

REDIS_PORT - Redis cloud port

Flask API endpoints

Send API key

  POST /key

Parameter	Type	Description
input	string	(Required) Your OpenAI API key

Send image

  POST /upload

Parameter	Type	Description
images	object (form-data)	(Required) Image to analyse

Tech Stack

Client: React, Redux, TailwindCSS

Server: Flask, Render, Redis

RAG: Pinecone, OpenAI models (gpt-4o, text-embedding-3-large), LangChain

todo:

1. link frontend to be processed by lc backend on flask ✔️
2. fix image processing errors - fix upload and api key endpoints ✔️
3. Convert image recognition results to embeddings to query pinecone vector store ✔️
4. prepare images for testing
5. firebase authentication
6. set up feedback database for RLHF