Description: This is a boilerplate for building genai applications. You will find examples of how to build applications in the following folders:
- fintech
- foodtech
- proptech
- similarity
- retail
Some of the tools and libraries used are:
- Datagen (for generating data) link: https://datagen.shadai.ai
- LangSmith (for tracing)
- LangGraph (for building agents)
- LangChain (for building assistants)
- ChromaDB (for vector search)
- FastAPI (for the api)
- Streamlit (for the ui)
- AWS Bedrock (for the llm)
- Claude LLM (for the llm)
Run the following command to start the application:
docker compose up -d --build
To see logs, run:
docker compose logs -f -tail 100 container_name
To restart the application, run:
docker compose restart
To stop the application, run:
docker compose down
Database is in postgres service and can be accessed with the following credentials:
postgres://admin:123asd456@localhost:5432/testdb
Redis is in redis service and can be accessed with the following route:
redis://localhost:6379
This api runs on port 8080 and can be accessed with the following route:
http://localhost:8080/
This UI with streamlit runs on port 8081 and can be accessed with the following route:
http://localhost:8081/
To create a person, send a POST request to the following endpoint:
http://localhost:8080/create-person
with the following body:
{
"email": "test@test.com",
"first_name": "John",
"last_name": "Doe"
}
For amazon bedrock, the credentials are:
AWS_DEFAULT_REGION=us-east-1
AWS_ACCESS_KEY_ID=ID
AWS_SECRET_ACCESS_KEY=KEY