🦜🔗 Langchain Experiments

This repo contains examples of using LangChain

📖 Contents

In particular, all main modules of LangChain are demonstrated in the notebooks.

1. 1_MODEL_IO.ipynb — Building blocks for interfacing with LLMs and Chat Models, using Prompt Templates and Output Parsers.
2. 2_MEMORY.ipynb — Using Memory buffers or summaries to store information during conversations, since LLMs are stateless.
3. 3_CHAINS.ipynb — Create more complex applications by connecting together multiple LLMs components into a Chain.
4. 4_AGENTS.ipynb — Create Agents that can leverage LLMs to interact with other Tools, from Wikipedia to Wolfram Alpha, from Google Search to Yahoo Finance.
5. 5_RETRIEVAL.ipynb — Implement Retrieval systems for running Question Answering over collections of documents or graph databases like knowledge graphs.

🛠️ Setup

First of all you need to define some variables for the environment that will be used in the notebooks. We do this with a .env file, so first copy the provided example .env.example to .env and fill in the values.

cp .env.example .env
# And now open ".env" with your favorite editor and fill in the values!

Then, you need to create a virtual environment and install the Python requirements.

python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Finally, if you want to run the Neo4j example (this is completely optional!!) in 5_RETRIEVAL.ipynb, you need to have a Neo4j instance running. To do so, you first need to install Docker, make sure that you have the Docker daemon running, and then run the following command.

docker run \
    --name neo4j \
    -p 7474:7474 -p 7687:7687 \
    -d \
    -e NEO4J_AUTH=neo4j/pleaseletmein \
    -e NEO4J_PLUGINS=\[\"apoc\"\]  \
    neo4j:latest