Skip to content

The project involves using the Wikipedia API to retrieve current content on a topic, and then using LangChain, OpenAI and Chroma to ask and answer questions about it. The aim of the project is to showcase the powerful embeddings and the endless possibilities.

Notifications You must be signed in to change notification settings

kolife01/chroma-langchain-sandbox

 
 

Repository files navigation

Chroma and LangChain Demo

This repository contains code and resources for demonstrating the power of Chroma and LangChain for asking questions about your own data. The demo showcases how to pull data from the English Wikipedia using their API. The project also demonstrates how to vectorize data in chunks and get embeddings using OpenAI embeddings model.

We then use LangChain to ask questions based on our data which is vectorized using OpenAI embeddings model. I used Chroma a database for storing and querying vectorized data.

Getting Started

To get started with the demo, you will need to have Python (I use Python 3.8) installed on your machine. You will also need to install the required Python packages by running the following command: pip install -r requirements.txt

You can change the _ALGORITHMS constant to whatever you want to query other topics on Wikipedia.

From there on you can simply run wikipedia.py which generates the text file which will be vectorized and stored in the database. You need to use the name of the created textfile in the ask_wikipedia.py file.

Now you can run ask_wikipedia.py.

Simply change the print(genie.ask("Can you tell me the formula for Linear Regression?")) in the ask_wikipedia.py file to whatever question you want to ask.

Video

I also created a video to demonstrate the demo. Screenshot

About

The project involves using the Wikipedia API to retrieve current content on a topic, and then using LangChain, OpenAI and Chroma to ask and answer questions about it. The aim of the project is to showcase the powerful embeddings and the endless possibilities.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%