Overview:

It is a Recipe Recommendation API (PoC) based on the supplied list of ingredients using RAG with generative AI.

Steps to run:

1. Unzip the project directory.
2. Set up the following env variables:

   export OPENAI_API_KEY=<your-openai-api-key>
   export RECIPE_URL=<url-to-recipe-dataset>
   export POSTGRES_URL=<postgres-connection-url>

3. Run the services using Docker Compose: docker-compose up

Sample Request:

1. Success Request:

   curl --location 'http://localhost:8000/recommend_recipe' \
   --header 'Content-Type: text/plain' \
   --data '"flour, sugar, baking powder, yogurt, egg"'

2. Failure Request:

   curl --location 'http://localhost:8000/recommend_recipe' \
   --header 'Content-Type: text/plain' \
   --data '"paper, ink"'

Assumptions & limitations

• Doesn't consider quantity
• Returns only one recipe
• Returns generic response even if the query is not correct.
• Does not implement the image based retrival.

Suggested Improvements

• Data ingestion should be separate from the app.
• Needs tests
• Error handling required.