Information Retrieval System

A text retrieval system implementing various information retrieval models with a Streamlit-based user interface.

Features

• Multiple retrieval models:
  • Boolean Retrieval Model
  • Vector Space Model (VSM)
  • Latent Semantic Analysis (LSA)
  • Combined Model (Boolean + Vector)
• Interactive search interface
• Document statistics and visualizations
• Model evaluation metrics

Installation

1. Clone this repository:

git clone https://github.com/dangvonguyen/IR-CS419.P21.git
cd IR-CS419.P21

2. Set up the environment and install dependencies

# Using pip
python -m venv .venv
source .venv/bin/activate
pip install -e .

# Using uv (faster installation)
uv sync
source .venv/bin/activate

Usage

Run the application using Streamlit:

streamlit run app.py

Using the Application

1. Load Data: Use the sidebar to select data source and parameters
2. Select Model: Choose between Boolean, VSM, LSA, or Combined retrieval models
3. Search: Enter queries in the search tab to retrieve relevant documents
4. Analyze: View document statistics and model performance in the Statistics tab
5. Browse: View loaded documents in the Documents tab

Project Structure

• app.py: Main Streamlit application
• src/models/: Implementation of retrieval models
  • boolean_model.py: Boolean retrieval with inverted index
  • vsm_model.py: Vector Space Model with TF-IDF
  • lsa_model.py: Latent Semantic Analysis model
  • combined_model.py: Combined Boolean and Vector model
• src/utils.py: Utility functions for text processing
• src/evaluate.py: Evaluation metrics for retrieval models
• ui/: User interface components