Mini Search Engine

Search engine in Python that builds a positional inverted index from .txt files and supports:

• ✅ Phrase search (e.g., "machine learning")
• ✅ BM25 ranking for relevance
• ✅ Snippet generation with highlighted query terms
• ✅ Command-line interface (CLI) powered by Typer

---

🚀 Features

• Indexing: Parse and normalize text, remove stopwords, and build a positional inverted index.
• Phrase Search: Supports exact multi-word phrases using a two-pointer merge algorithm.
• Ranking: Implements BM25 scoring for relevant document ranking.
• Snippets: Extracts text fragments around matches and highlights query terms.
• CLI: Easy-to-use interface with two commands: index and query.

---

📂 Project Structure

Mini Search Engine/
  mini_search/
    __init__.py
    cli.py
    indexing.py
    planner.py
    phrase.py
    ranking.py
    snippets.py
    tokenize.py
    types.py
  documents/         # Example .txt documents
  README.md

---

⚙️ Installation

Clone the repo and install dependencies:

pip install typer[all]

---

📥 Index Documents

Put .txt files into the documents/ folder.
Then build the index:

python -m mini_search.cli index documents/ -o index.pkl

---

🔎 Search

Run a query:

python -m mini_search.cli query python -i index.pkl

Phrase queries are supported:

python -m mini_search.cli query python "machine learning" -i index.pkl -k 5

Example output:

Top results:
- doc2.txt   BM25: 5.8421
  ...Machine [learning] is a field of artificial intelligence that uses statistical techniques...

- doc5.txt   BM25: 4.9130
  ...Deep [learning] architectures such as deep neural networks...

---

🛠 Tech Stack

• Python 3.10+
• Typer for CLI
• Dataclasses and Type Hints for clean code
• Pickle for index persistence

---