README.md

metadata

title: Simple Search Engine
emoji: 🔍
colorFrom: blue
colorTo: purple
sdk: docker
app_port: 7860
pinned: false

Simple Search Engine 🔍

An intelligent document search engine powered by Sentence Transformers (SBERT) and FastAPI.

Features

• Semantic Search: Uses the all-MiniLM-L6-v2 model for understanding query context
• Fast & Efficient: Built with FastAPI for high performance
• Beautiful UI: Clean, modern interface with gradient design
• Real-time Results: Instant search results with similarity scores

How It Works

1. Documents are chunked into smaller segments (3 sentences each)
2. Each chunk is encoded using SBERT into vector embeddings
3. User queries are encoded and compared using cosine similarity
4. Top 5 most relevant chunks are returned with similarity scores

Technology Stack

• Backend: FastAPI
• ML Model: Sentence Transformers (all-MiniLM-L6-v2)
• NLP: NLTK for sentence tokenization
• Similarity: Scikit-learn for cosine similarity computation

Usage

Simply enter your search query in the search box and press Enter or click the Search button. The engine will return the top 5 most relevant document chunks with their similarity scores.

API Endpoints

• GET / - Web interface
• POST /search - Search endpoint (accepts JSON with query field)
• GET /health - Health check endpoint

Example Queries

• "machine learning AI"
• "cloud infrastructure AWS"
• "financial reports revenue"
• "marketing SEO strategies"