Document Searcher

A desktop application for searching through Excel, Word, and PDF files. Built with Electron, React, and Python.

Features

• Multi-format search -- Search across .xlsx, .xls, .docx, and .pdf files
• Multi-folder search -- Select multiple folders to search at once
• Tabbed interface -- Run independent searches in separate tabs (Ctrl+T / Ctrl+W)
• Preview pane -- Click a result to see surrounding content in context
• Search history -- Recent searches are saved and accessible from a dropdown
• Search indexing -- SQLite-backed persistent index for faster repeat searches
• CSV export -- Export filtered results to CSV
• Settings -- Configure file types, indexing, max results, auto-update preferences
• Auto-update -- Checks GitHub Releases for new versions
• Multi-window -- Open additional windows with Ctrl+Shift+N
• Cross-platform -- Builds for Windows (NSIS), macOS (DMG), and Linux (AppImage/DEB)

Project Structure

main/
  main.js          -- Electron main process (windows, IPC, settings, auto-update)
  preload.js       -- Context bridge API for renderer
src/
  main.jsx         -- React entry point
  App.jsx          -- App shell with tab management
  components/      -- React components (SearchTab, ResultsTable, PreviewPane, etc.)
  hooks/           -- Custom hooks (useSearch, useSettings, useHistory)
  styles/          -- Global CSS (dark theme)
python/
  search.py        -- Search backend (Excel, Word, PDF, indexing, preview)
  indexer.py        -- SQLite persistent index module
  requirements.txt -- Python dependencies

Development

Prerequisites

• Node.js 18+
• Python 3.9+
• pip

Setup

npm install
pip install -r python/requirements.txt

Run in development

npm run dev

This starts the Vite dev server and Electron concurrently.

Build installer

npm run build

Output goes to dist_electron/.

Build and publish a release

$env:GH_TOKEN = (gh auth token)
npm run build -- --publish always

Python Backend

The Python backend communicates with Electron via newline-delimited JSON over stdin/stdout. It supports these commands:

Command	Description
search	Search files in folders for a query string
preview	Extract surrounding context for a matched result
index	Build/refresh the SQLite search index
search_indexed	Search using the persistent index
index_status	Get indexed file count and staleness info
clear_index	Delete the index database

Bundling for production

Use PyInstaller to freeze the Python backend:

pyinstaller search.spec

Place the output binary in resources/ before running npm run build.

License

MIT