Project

General

Profile

Actions

dev #1578

open

Build a Demo Search Engine Algorithm for the App

Added by Zahid Hassan over 1 year ago. Updated about 1 year ago.

Status:
Complete
Priority:
High
Assignee:
Target version:
-
Start date:
01/13/2025
Due date:
01/13/2025 (about 15 months late)
% Done:

100%

Estimated time:
32:00 h
Spent time:
27:00 h (Total: 30:00 h)

Description

Description:

Develop a prototype search engine algorithm to be integrated into the app. The algorithm should efficiently retrieve relevant results based on user queries, providing a solid foundation for future enhancements.
The demo will serve as a proof of concept and focus on accuracy, speed, and scalability.

Objectives:

  • Search Functionality: Implement basic search capabilities that allow users to query data using keywords.
  • Result Ranking: Design an algorithm to rank search results based on relevance to the query.
  • Indexing: Create a lightweight indexing mechanism for efficient data retrieval.
  • Performance Optimization: Ensure the algorithm performs well with a small to medium dataset ( *~ 5,000 to 10,000 records* ).

Key Features:

  • Full-text Search: Enable keyword-based search on the dataset.
  • Relevance Scoring: Rank results using basic relevance scoring ( term frequency-inverse document frequency (TF-IDF) ).
  • Error Handling: Handle cases where no results are found or when invalid queries are provided.
  • Suggestions: Provide basic query suggestions for misspelled words.

Tasks:

  • Analyze requirements and define the structure of the dataset to be used.
  • Data pre-processing and indexing.
  • Query parser to interpret user input.
  • Search algorithm for retrieving and ranking results.
  • Integrate the search engine with the app's existing backend.
  • Build a minimal console interface to test the demo search functionality.
  • Test the search engine for accuracy and speed with test cases.
  • Document the algorithm, functionality, and setup instructions for further improvements.

Acceptance Criteria:

  • Users can perform searches and receive relevant results.
  • Results are ranked based on keyword relevance.
  • The system handles basic edge cases ( empty queries, misspelled keywords ).
  • A console demo is available for testing the search engine.

Subtasks 1 (1 open0 closed)

dev #1662: Response time analysis and context based searchComplete Zahid Hassan01/13/202501/13/2025

Actions
Actions #1

Updated by Zahid Hassan about 1 year ago

  • Subtask #1662 added
Actions #2

Updated by Zahid Hassan about 1 year ago

  • Status changed from In Progress to Complete
Actions

Also available in: Atom PDF