Focus Group RAG Summariser

This project provides a Retrieval-Augmented Generation (RAG) system for summarizing and analyzing focus group transcripts. The system uses a combination of LangChain, Ollama's Llama model, and a Chroma vector store to retrieve relevant information and generate summaries based on the input data.

Features

• Upload Focus Group Data: Upload CSV files containing focus group transcripts.
• Query and Summarization: Ask questions related to the topics of the transcripts and receive relevant summaries.
• Feedback Collection: Collect user feedback on responses to improve the system's accuracy.
• Visualization: Visualize the collected feedback with interactive charts.

Requirements

To run the project, you'll need the following dependencies. Install them using:

pip install -r requirements.txt

Setup Instructions

1. Clone the repository:

git clone <repository-url>
cd <project-folder>

2. Create a virtual environment:

python3 -m venv venv

3. Activate the virtual environment:

• On MacOS/Linux:

source venv/bin/activate

• On Windows

venv\Scripts\activate

4. Install dependencies:

pip install -r requirements.txt

5. Run the Streamlit app:

streamlit run focus_group_summariser_ollama_feedback_app.py

This will launch the app in your web browser.

Usage

• Upload CSV: Upload a CSV file containing your focus group transcripts with columns transcript and topic_name.
• Select a Topic: Choose a topic from the dropdown (optional).
• Ask a Question: Type in your question or request a summary.
• Get Feedback: After receiving an answer, you can rate the response and leave feedback.
• View Collected Feedback: Toggle the visibility of feedback and view related statistics in the last 4 weeks.

Feedback

After interacting with the app, users can provide feedback on the responses to improve the system. Feedback is stored in a CSV file for further analysis and visualization.

Technologies Used

• Streamlit: For building the web interface.
• LangChain: For integrating language models with the RAG system.
• Ollama: For the Llama model powering the summarization process.
• Chroma: A vector store for efficient retrieval of information.
• Matplotlib: For feedback visualization.

Contributing

Feel free to fork this project and submit issues or pull requests. Contributions are welcome!

Notes

This app requires the download and local installation/hosting of Ollama in order to use the Llama model. Readers can access these models for free by visiting https://ollama.com/.