Positional Inverted Index

Overview

This project implements a Positional Inverted Index in Python, enabling efficient document retrieval based on term proximity. The InvertedIndex class indexes a document corpus and provides search functionality for queries with two terms separated by a specified distance (k). The index supports bidirectional and unidirectional positional proximity search algorithms.

Developed for CSC 583 under Instructor Mihai Surdeanu.

Features

• Positional Indexing: Stores term positions for efficient proximity-based retrieval.
• Bidirectional and Unidirectional Search:
  • Bidirectional Search: Finds terms within a distance k of each other in either direction.
  • Unidirectional Search: Finds terms within a distance k with the first term preceding the second.
• Configurable Proximity Queries: Handles queries in the format term1 /k term2 to specify the maximum distance between terms.

Public Methods

read_txt_file(input_file: str) -> list: Reads the document file line by line, returning non-empty lines.

• q7_1_1(query: str) -> list: Executes a bidirectional proximity query.
• q7_1_2(query: str) -> list: Executes a second bidirectional proximity query.
• q7_2(query: str) -> list: Executes a unidirectional proximity query.