LangChain Documentation Helper

Overview

The LangChain Documentation Helper, built with Langchain, pinecone and streamlit, is a web-based application designed to facilitate learning and exploration of LangChain by leveraging generative AI. By integrating with Pinecone, a vector database, this application provides an efficient and intuitive way to query and interact with LangChain's official documentation.

Environment Setup

To securely manage application configurations and sensitive credentials, this project uses environment variables stored in a .env file at the root of the project directory. Before running the application, you must create this file and populate it with the necessary keys.

env File Contents

Your .env file should contain the following key-value pairs:

OPENAI_API_KEY='your_openai_api_key_here'
PINECONE_API_KEY='your_pinecone_api_key_here'
PINECONE_ENVIRONMENT_REGION='your_pinecone_region_here'

Please ensure that your .env file is never committed to your version control system. Add .env to your .gitignore file to prevent accidental uploads of sensitive information.

Run locally with Docker

Setup

Dockerfile

# Base image selection
FROM python:3.11.7-bookworm

# Set working directory in container
WORKDIR /app

# Set up networking to container
EXPOSE 8501

# Copy local files to the container location
COPY . /app/

# Set up environment
# RUN mkdir /app/langchain-docs
RUN apt update && apt install -y wget
RUN pip install pipenv
# Download langchain documentation
RUN wget -r -A.html -P ./langchain-docs https://api.python.langchain.com/en/latest/index.html

# Set up dependencies for the documentation helper flask server
RUN pipenv --python /usr/local/bin/python
RUN pipenv install

# Upload langchain docs to Pinecone
RUN pipenv python setup_ingestion.py

# Container start command
CMD [ "pipenv", "run", "streamlit", "run", "app.py" ]

1. Clone the repository:

git clone https://github.com/SueGK/documentation-helper.git

2. Navigate to the project directory:

cd documentation-helper

3. Run the Script to build and run docker container

python start-doc-helper.sh  

start-doc-helper.sh contains:

#!/bin/bash -x

# Builds the local image
docker build -t documentation-helper:0.1 .

# Start the container
docker run -d --name docs-h -p 8501:8501 docs-helper:0.1

🥳 Now, go to http://localhost:8501/ to talk with your bot!

Deploy to streamlit cloud

🔥 CLICK IMAGE TO START TALKING TO THE BOT, SET UP YOUR API AT API SETTING PAGES

Setup

1. Clone the repository:

git clone https://github.com/SueGK/documentation-helper.git

2. Navigate to the project directory:

cd documentation-helper

3. Change to deploy-streamlit-cloud branch

git checkout deploy-streamlit-cloud

4. Install dependencies

pip install -r requirements.txt

5. Download LangChain Documentation

  mkdir langchain-docs
  wget -r -A.html -P langchain-docs  https://api.python.langchain.com/en/latest

6. Embed LangChain Official Documentation into Pinecone Index Using OpenAI's Embedding Model

python setup_ingestion.py

7. Start the server

streamlit run app.py

8. Deploy on streamlit cloud

Debug Streamlit Applications in VSCode

To debug a Streamlit application, you need to create a launch configuration for Python:

Go to the Run and Debug view by clicking on the play icon with a bug on the sidebar or pressing Ctrl+Shift+D (Cmd+Shift+D on macOS). Click on "create a launch.json file" link, then select the Python environment you're using for your Streamlit app. Replace the content of the launch.json file with the following configuration:

{
    "version": "0.2.0",
    "configurations": [
        
        {
            "name": "Python Debugger: Current File",
            "type": "debugpy",
            "request": "launch",
            // use pipenv --venv to change the path
            "program": "PATH/TO/VENV/bin/streamlit",
            "console": "integratedTerminal",
            "justMyCode": true,
            "args": ["run","app.py"]
        }
    ]
}

Make sure to change "program": "PATH/TO/VENV/bin/streamlit"