TODO
Client: Streamlit
Server Side: LangChain 🦜🔗
Vectorstore: Pinecone 🌲
To run this project, you will need to add the following environment variables to your .env file
PINECONE_API_KEY
OPENAI_API_KEY
LANGCHAIN_TRACING_V2
LANGCHAIN_API_KEY
LANGCHAIN_PROJECT
PROJECT_PROFILE_HELPER_PASSWORD
Clone the project
git clone git@github.com:cwmat/response-search.git
See https://github.com/emarco177/documentation-helper.git
for more context, forked from this repo.
Go to the project directory
cd documentation-helper
Download Docx data and place in
./data/
Install dependencies
pipenv install
Push data to Pinecone (need to create the index first and update the name in ingestion.py
and backend/core.py
)
Run ingestion script to hydrate Pinecone index with docx embeddings
pipenv run ingest
Start the streamlit server
pipenv run start