fastapi-langchain

Showcase pydantic, fastapi, langchain, openai

live web app running on: http://54.198.111.198:8080/static/index.html

we use pydantic to validate:

• all rows going into pinecone have data
• all pinecone data cites a valid https link
• the validation is done before the upsert
• the data used here was validated in scripts/pydantic_validation_langchain_retrieval_augmentation.ipynb
  • they were only done once for upserting because once it is validated as a dataset before going into pinecone, we dont need to validate before upserting it again
  • we would validate again if we were refreshing our pinecone embeddings with new data

to run inside ec2 for demo purposes (if production would use nginx): uvicorn backend.api.main:app --host 0.0.0.0 --port 8080

• to run locally, use command: uvicorn backend.api.main:app --reload and check page http://127.0.0.1:8000/static/index.html

  to restart the local instance: lsof -i :8000 kill -9 <pid>

file structure

fastapi-langchain/
│ ├── backend/ │ ├── data/ │ │ └── processed/ │ │ └── documents/ │ │ └── parquet_df_with_metadata.parquet │ │ │ ├── scripts/ │ │ ├── live_LLM.py │ │ ├── ... [other scripts] │ │ │ ├── api/ │ │ ├── main.py # FastAPI main application file │ │ ├── models/ # Pydantic models, if any │ │ └── routers/ # Separate FastAPI routers, if you want to modularize endpoints │ │ │ └── .env # Contains environment variables │ ├── frontend/ │ └── requirements.txt