LLM for PyHealth

The current LLM for pyhealth interface is deployed here: http://35.208.88.194:7861/.

Step 1:

Merge all pyhealth related information (code and doc txt) into pyhealth.txt

• currently, we use https://github.com/mpoon/gpt-repository-loader

Step 2:

Run the ingest.py to transform the pyhealth.txt into the FAISS vector database

python ingest.py

Step 3:

Run the retrieval augmented generation (RAG) app for Q & A based on the pyhealth.txt document.

python app_rag.py

Launch in Docker

1. Modfiy environment variables (OPENAI_API_KEY, server address...)in Dockerfile.
2. Build image by docker build -t chat-pyhealth ..
3. Debug a container by docker run -p 0.0.0.0:7861:7861 --name chat-pyhealth-c -v ./logs/:/app/logs/ chat-pyhealth.
4. Run a container by docker run -d -p 0.0.0.0:7861:7861 --name chat-pyhealth-c -v ./logs/:/app/logs/ chat-pyhealth.

## build container
docker run -d -p [host address and port]:[container port] --name [name] -v [host path]:[container path] [image]

# -d: detached
# -p: port
# --name: container name
# -v: mount directory of container to local host path

## check
docker ps
docker images

## remove
docker stop / restart chat-pyhealth-c # container
docker rm chat-pyhealth-c # container
docker rmi chat-pyhealth # image

## modify directly
docker cp [local file in host] chat-pyhealth-c:[container path]

Let me know if you want to join and help us improve the current interface.