https://github.com/shruagarwal/activeloop-langchain-coursebot

Ask questions related to langchain course brought by Activeloop.
https://github.com/shruagarwal/activeloop-langchain-coursebot

apify apify-api cohere langchain llm-chatbot openai-api streamlit

Last synced: 3 months ago
JSON representation

Ask questions related to langchain course brought by Activeloop.

• Host: GitHub
• URL: https://github.com/shruagarwal/activeloop-langchain-coursebot
• Owner: ShruAgarwal
• Created: 2024-03-19T12:34:38.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2024-04-15T12:56:59.000Z (about 1 year ago)
• Last Synced: 2024-04-21T10:22:21.290Z (about 1 year ago)
• Topics: apify, apify-api, cohere, langchain, llm-chatbot, openai-api, streamlit
• Language: Python
• Homepage: https://educhain-bot.streamlit.app/
• Size: 10.7 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# 🤖 Chat with EduChain Bot

### An LLM companion for answering your questions related to the [LangChain & Vector DBs in Production course](https://learn.activeloop.ai/courses/langchain) brought by *[Activeloop](https://www.activeloop.ai/)*

## Demo 🕹

[![Streamlit App](https://static.streamlit.io/badges/streamlit_badge_black_white.svg)](https://educhain-bot.streamlit.app/)

https://github.com/ShruAgarwal/Activeloop-LangChain-CourseBot/assets/82811717/55c29e11-cc27-4e9d-8d69-848a48e724e8

## How to use 👀

1. Enter your OpenAI API key.

   - You can get your own OpenAI API key from [here](https://platform.openai.com/account/api-keys) and then click on the `+ Create new secret key` button.

2. You can now proceed to ask questions related to the course to the chatbot.

## Behind the Scenes ⚙

This educational chatbot demonstrates the power of **Retrieval Augmented Generation (RAG)** to answer queries related to the course and provides relevant info to you by retrieving data from an extensive and detailed knowledge base. It returns a natural response to your questions along with the truth source.

*Here's a summary of the scripts used for building this chatbot:*

1. `data_loading.py`:

   - Handles the initial data gathering and processing task where it scrapes the text data from the Langchain course website using **[Apify](https://apify.com/)**.

   - The scraped text data is then converted into numerical form (vectors) using [`CohereEmbeddings`](https://docs.cohere.com/docs/embeddings) that the chatbot can learn from.

   - Finally, the transformed data is uploaded to **[Deep Lake](https://docs.activeloop.ai/)**, a data storage service, for future use.

2. `app_workflow.py`:

   - Handles the retrieval and ranking of the relevant data.

   - First, it gathers the stored data from the `data_loading.py` file.

   - Second, it uses `CohereRerank` to rank and retrieve the most relevant data based on the user’s query.

      - `CohereRerank` is a reranking service that refines and ranks documents in alignment with a user’s search criteria.

   - Third, it also builds the conversation chain with memory, which helps in maintaining the context of the conversation.

3. `app_demo.py`:

   - The main script demonstrates the working of the chatbot through a user-friendly web interface using **Streamlit**.

   - The chatbot then uses the stored and transformed data from **Deep Lake** to answer user queries.

### Tech-stack 🛠



  



## Key Learnings 🌱

- Provided a deep understanding of how RAG can be used to answer queries by retrieving relevant information from a detailed knowledge base.

- Involves scraping data from a course website, which helped in understanding how to extract and structure data from the web.

- The use of **Cohere** for embedding and reranking provided insights into how these techniques can improve the relevance of the retrieved information.

- Storing the transformed data in Deep Lake helped in understanding the importance of efficient data storage and retrieval in AI applications.

- Demonstrated how to integrate powerful APIs and libraries like **OpenAI and Langchain** to build a sophisticated chatbot.

- Highlightes the potential of AI in enhancing educational experiences, by providing a chatbot that can answer course-related queries.

- Helped with learning *project management, problem-solving, and debugging skills* to bring all the components together into a working chatbot.

## Credits ✨

- *Inspired to build this chatbot from* [this tutorial!](https://www.activeloop.ai/resources/retrieval-augmented-generation-for-llm-bots-with-lang-chain/)

- *Thanks to* [Yuichiro's Streamlit Theme Editor](https://github.com/whitphx/streamlit-theme-editor) that helped me find the suitable app's theme :)

- *Chatbot logo and tech stack design made by me using [Canva](https://www.canva.com/)*