Skip to content

Chatbot with Chainlit#

This example demonstrates a simple chatbot using Quivr and Chainlit, where users can upload a .txt file and ask questions based on its content.

Prerequisites#

  • Python: Version 3.8 or higher.
  • OpenAI API Key: Ensure you have a valid OpenAI API key.

Installation#

  1. Clone the repository and navigate to the appropriate directory: 

    git clone https://github.com/QuivrHQ/quivr
cd examples/chatbot

  2. Set the OpenAI API key as an environment variable: 

    export OPENAI_API_KEY='<your-key-here>'

  3. Install the required dependencies: 

    pip install -r requirements.lock

Running the Chatbot#

  1. Start the Chainlit server: 

    chainlit run main.py

  2. Open your web browser and navigate to the URL displayed in the terminal (default: http://localhost:8000).

Using the Chatbot#

File Upload#

  1. On the chatbot interface, upload a .txt file when prompted.
  2. Ensure the file size is under 20MB.
  3. After uploading, the file is processed, and you will be notified when the chatbot is ready.

Asking Questions#

  1. Type your questions into the chat input and press Enter.
  2. The chatbot will respond based on the content of the uploaded file.
  3. Relevant file sources for the answers are displayed in the chat.

How It Works#

  1. File Upload:

    • Users upload a .txt file, which is temporarily saved.
    • The chatbot processes the file using Quivr to create a "brain."

  2. Session Handling:

    • Chainlit manages the session to retain the file path and brain context.

  3. Question Answering:

    • The chatbot uses the ask_streaming method from Quivr to process user queries.
    • Responses are streamed incrementally for faster feedback.
    • Relevant file excerpts (sources) are extracted and displayed.

  4. Retrieval Configuration:

    • A YAML file (basic_rag_workflow.yaml) defines retrieval parameters for Quivr.

Workflow#

Chat Start#

  1. Waits for the user to upload a .txt file.
  2. Processes the file and creates a "brain."
  3. Notifies the user when the system is ready for questions.

On User Message#

  1. Retrieves the "brain" from the session.
  2. Processes the user's question with Quivr.
  3. Streams the response and displays it in the chat.
  4. Extracts and shows relevant sources from the file.

Features#

  1. File Processing: Creates a context-aware "brain" from the uploaded file.
  2. Streaming Responses: Delivers answers incrementally for better user experience.
  3. Source Highlighting: Displays file excerpts relevant to the answers.

Enjoy interacting with your text files in a seamless Q&A format!