Skip to content

Latest commit

 

History

History
149 lines (104 loc) · 4.7 KB

README.md

File metadata and controls

149 lines (104 loc) · 4.7 KB

Localume

Python Version: 3.8+ OS: Windows 11 Version: v1.0.0

Localume A modern desktop application for semantic file search and real-time monitoring using vector embeddings and LLM-powered search optimization.

Project Bunner

Description

Localume is a powerful desktop application that enables semantic search across your documents using advanced vector embeddings and retrieval technology. The application monitors specified directories in real-time, automatically indexing new and modified files to maintain an up-to-date searchable database.

Key Features

  • Semantic Search: Find documents based on meaning, not just keywords

  • Real-time Monitoring: Automatic indexing of new and modified files

  • Multiple File Support: Handles PDF, TXT, and Markdown files

  • Modern UI: Clean, intuitive interface with light/dark theme support

  • System Tray Integration: Runs quietly in the background

  • LLM-Powered: Query optimization using Google's Gemini API (Optional)

Installation

Prerequisites

  • Python 3.8+

  • Windows 11 (64-bit) (tested on windows for the moment)

  • Google Gemini API key(Optional)

Setup

  1. Clone the repository:
git clone https://github.com/yourusername/smart-file-search.git
cd smart-file-search
  1. Create and activate a virtual environment:
python -m venv venv
.\venv\Scripts\activate
  1. Install dependencies:
pip install -r requirements.txt

Usage

  1. Start the application:
cd ui
python3 gui_app.py

  1. Add folders to monitor:

    • Click "Add Folder" button
    • Select directories containing documents to index
    • The application will begin scanning and indexing files

  2. Search for documents:

    • Enter your search query in natural language
    • Results will be displayed with relevant metadata
    • Double-click results to open files

Code Examples

The application uses a modular architecture. Here's an example of performing a search:

from core.search.search_engine import SearchEngine
from core.embeddings.embedding_generator import EmbeddingModel
from core.embeddings.vector_store import VectorStore

# Initialize components
vector_store = VectorStore(dimension=384)
embedding_model = EmbeddingModel()
search_engine = SearchEngine(vector_store, embedding_model)

# Perform search
results = search_engine.search("documents about project requirements", top_k=10)

Architecture

The project follows a clean, modular architecture:

smart-file-search/
├── core/
│   ├── embeddings/     # Vector embedding generation and storage
│   ├── scanner/        # File monitoring and content extraction
│   ├── search/         # Search engine implementation
│   ├── llm/            # LLM service integration
│   └── utils/          # Shared utilities
│
├── data/
│   ├── faiss.index     # stores document vector embeddings
│   └── id_map.pkl      # stores a mapping between document IDs and their metadata
│
├── ui/                 # Desktop GUI implementation

Contributing

Contributions are welcome!

Development Guidelines

  • Follow PEP 8 style guide
  • Add type hints to all functions
  • Include docstrings for classes and methods
  • Write unit tests for new features
  • Update documentation as needed

License

This project is licensed under the GNU General Public License (GPL-3.0) License - see the LICENSE file for details.

Acknowledgments

  • Azure Theme for the modern UI
  • sentence-transformers for text embeddings
  • Google's Gemini API for query optimization
  • Facebook vector database FAISS

Contact

- Mohamed Karim Ben Boubaker | LinkedIn 

FAQ

Q: How does the semantic search work?

A: The application converts documents into vector embeddings using sentence transformers, enabling similarity-based search that understands context and meaning rather than just matching keywords.

Q: What file types are supported?

A: Currently, the application supports PDF, TXT, and Markdown files. Support for additional file types is planned for future releases.

Note: This project is under active development. Features and documentation may be updated frequently.