IITD Campus Navigator: Vector Search System

Version 1.0 | January 2025

Overview

IITD Campus Navigator is a chat-based system that helps users explore and understand life at IIT Delhi. Using vector search and language models, it provides information about:

• Academic programs and courses
• Campus facilities and navigation
• Student life and experiences
• Career guidance and opportunities

System Architecture

Core Components

[Streamlit Frontend] <-> [Chat Interface] <-> [Vector Search] <-> [Qdrant]
                                        |
                                  [Groq LLM API]

Technology Stack

• Frontend: Streamlit
• Vector Database: Qdrant
• Embeddings: SentenceTransformer (all-MiniLM-L6-v2)
• LLM: Groq API
• Environment: Poetry, Nix

Data Sources

Academic Information

1. Course Catalog

   • Scraped course details
   • Course codes and names
   • Credit information
   • Prerequisites
   • Course descriptions

2. BSP Academic Resources

   • Course experiences
   • Study materials
   • Academic guidance
   • Student testimonials
   • Intern Fundae

Inception Magazine Content

The freshers' magazine "Inception" provides:

• Campus facility descriptions
• Hostel information
• Navigation guides
• First-year experiences
• Essential campus knowledge

Dalle United Magazine Content

BSP's "Dalle United" magazine contributes:

• Student interviews
• Campus life stories
• Cultural coverage

Vector Store Implementation

Data Processing

1. Text Preprocessing

   • Cleaning and formatting
   • Context preservation
   • Campus terminology handling

2. Vector Generation

   • Text-to-vector conversion
   • Semantic embedding creation
   • Context maintenance

3. Collection Structure

   • Organized by source type
   • Optimized for retrieval
   • Contextual relationships preserved

Development Setup

Environment Configuration

# Initialize development environment
nix develop
poetry install

# Start application
poetry run streamlit run app.py

Project Structure

pegasus/
├── LICENSE
├── README.md
├── app/app.py
├── final_courses.json
├── flake.lock
├── flake.nix
├── poetry.lock
├── pyproject.toml
└── scripts/ (all the misc scripts)

Query Processing

Search Flow

1. Query Analysis

   • Intent recognition
   • Context identification
   • Source prioritization

2. Response Generation

   • Context-aware answers
   • Source combination
   • Natural language output
   • Session aware

Planned Enhancements

Near-term

1. Content Updates

   • Regular magazine updates
   • New course information
   • Fresh student experiences
   • Add departmental data

2. Search Features

   • Better context handling
   • Multi-topic queries
   • Chat history

Long-term

1. User Experience

   • Mobile interface
   • Response speed
   • Better error handling
   • Streaming the text generated for a fluid exp

2. Data Coverage

   • More student resources
   • Additional magazines
   • Updated course content

Contributing

To contribute:

1. Fork the repository
2. Create a feature branch
3. Make changes with documentation
4. Submit a pull request

License

MIT License. See LICENSE file.

---

For detailed implementation information, refer to the scripts/ directory and development notebooks.