QueryVista: News Article Answerer and Ranker with RAG Pipeline (Project Under The Course - Natural Language Processing)

Demo - https://youtu.be/BzouxsT9nVQ QueryVista is a powerful application that summarizes and ranks news articles using a Retrieval-Augmented Generation (RAG) pipeline. By leveraging LangChain, ChromaDB, and Streamlit, this app allows users to input news article URLs, store them in a vectorized database, and retrieve relevant answers based on user queries.

Features:

• RAG-based Summarization: Uses a combination of retrieval and generation techniques to answer and rank articles.
• Vector Database: Stores articles as vectorized embeddings using ChromaDB for fast and efficient retrieval.
• Streamlit Interface: A simple web interface for users to input article URLs and query for relevant summaries.
• Summarization: Uses a free text summarization model from Hugging Face to summarize articles.
• Ranking: Ranks the articles based on relevance to the user’s query.

Technologies:

• LangChain: For orchestrating the RAG pipeline and managing the chain of operations.
• ChromaDB: For storing and managing article embeddings in a vector database.
• Streamlit: For building the user interface.
• Hugging Face: For using pre-trained models like text summarizers.

Setup Instructions:

Prerequisites:

Before you begin, make sure you have the following installed:

• Python (>= 3.8)
• pip (Python package installer)
• Docker (for containerized deployment, optional but recommended)

2. Install Dependencies:

pip install -r requirements.txt

3. Running the Application:

To run the application locally using Streamlit, simply execute:

streamlit run app/main.py

This will start the app and open it in your default web browser.

4. (Optional) Docker Deployment:

If you prefer to deploy the application in a Docker container, follow these steps:

1. Build the Docker image:

   docker build -t queryvista .

2. Run the Docker container:

   docker run -p 8501:8501 queryvista

The app will be accessible at http://localhost:8501.

How to Use:

1. Input News Articles: On the homepage, you can input URLs of news articles. The application will scrape and process the content, storing it in the vector database.

2. Query for Summaries: Once articles are stored, you can enter a query related to any of the stored articles. The app will rank the articles based on their relevance to the query and answer your query for you.

3. View Summaries: The summarized content of the top-ranked articles will be displayed on the screen. You can view the original article or the generated answer directly.

Code Structure:

QueryVista/
├── main.py                # Streamlit app interface
├── scraper.py             # Article scraping logic
├── vector_store.py        # Vector store and database operations
├── rag_pipeline.py        # Retrieval-Augmented Generation (RAG) pipeline
├── requirements.txt       # Required Python libraries
├── Dockerfile             # Dockerfile for containerizing the app
├── README.md   
├── LICENSE 
demo/
├── sample_input.txt       # Contains the sample/test case input to the app