Open source PDF reader built with Rust, Dioxus and PDFium

3 months ago 7

A modern PDF viewer application built with Rust and Dioxus, featuring AI-powered search capabilities using Gemini.

Dynamic File Selection: Open and close PDF files through an intuitive file dialog interface
High-Quality PDF Display: Renders PDF pages at 1000x1400 resolution for balanced quality and performance
Continuous Scroll: View all PDF pages in a seamless continuous scroll interface
AI-Powered Search: Integrated multi-provider AI API (Gemini, ChatGPT, Claude) for intelligent content analysis and term explanations
Flashcard System: Automatic vocabulary flashcard creation from search terms with persistent storage and popup interface
Reading Bookmark System: Page-level bookmark functionality to track reading progress across sessions
Position Marker System: Precise in-page position markers with automatic scrolling navigation
PDF Page Rotation: Individual page rotation with persistent state across sessions
Recent Files History: Quick access to recently opened PDF files (up to 10 files)
API Key Management: Automatic saving and restoration of AI provider API keys
Modern UI: Clean, responsive interface built with Dioxus framework
Optimized Layout: Specially designed for vertical PDF documents with horizontal AI search panel
Smart Caching: Intelligent page caching system with optimized loading to prevent unnecessary re-rendering
Cross-platform: Desktop application with native performance

Dioxus - Modern reactive UI framework for Rust
pdfium-render - PDF rendering using Google's PDFium library
rfd - Native file dialog integration for file selection
Reqwest - HTTP client for AI search functionality
Tokio - Async runtime for non-blocking operations
Serde - JSON serialization for API communication
Base64 - Image encoding for web display

This application requires Rust and Cargo. Install them from the official website:

Windows, macOS, and Linux:

# Install Rust via rustup (recommended) curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh # On Windows, download and run rustup-init.exe from: # https://rustup.rs/ # Restart your terminal or run: source ~/.cargo/env # Verify installation cargo --version rustc --version

Alternative Installation Methods:

macOS (via Homebrew):

Ubuntu/Debian:

sudo apt update sudo apt install rustc cargo

CentOS/RHEL/Fedora:

# CentOS/RHEL sudo yum install rust cargo # Fedora sudo dnf install rust cargo

To use the AI search functionality, obtain one or more of the following API keys:

Gemini API (Google): https://makersuite.google.com/app/apikey

ChatGPT API (OpenAI): https://platform.openai.com/api-keys

Claude API (Anthropic): https://console.anthropic.com/

# Clone the repository git clone <repository-url> cd rust-pdf-viewer # Create lib directory for PDFium library mkdir -p lib # Download PDFium library for Intel Macs (x64) curl -L -o pdfium-mac-x64.tgz \ "https://github.com/bblanchon/pdfium-binaries/releases/latest/download/pdfium-mac-x64.tgz" && \ tar -xzf pdfium-mac-x64.tgz -C lib --strip-components=1 && \ rm pdfium-mac-x64.tgz # For Apple Silicon Macs (ARM64), use this instead: # curl -L -o pdfium-mac-arm64.tgz \ # "https://github.com/bblanchon/pdfium-binaries/releases/latest/download/pdfium-mac-arm64.tgz" && \ # tar -xzf pdfium-mac-arm64.tgz -C lib --strip-components=1 && \ # rm pdfium-mac-arm64.tgz # Verify the library is correctly placed ls -la lib/libpdfium.dylib # Build and run the project cargo run # To build a macOS app bundle (.app file) cargo install cargo-bundle cargo bundle --release # The app bundle will be created at: # target/release/bundle/osx/PDF Viewer.app

# Clone the repository git clone <repository-url> cd rust-pdf-viewer # Create lib directory for PDFium library mkdir -p lib # Download PDFium library for x64 systems curl -L -o pdfium-linux-x64.tgz \ "https://github.com/bblanchon/pdfium-binaries/releases/latest/download/pdfium-linux-x64.tgz" && \ tar -xzf pdfium-linux-x64.tgz -C lib --strip-components=1 && \ rm pdfium-linux-x64.tgz # For ARM64 systems, use this instead: # curl -L -o pdfium-linux-arm64.tgz \ # "https://github.com/bblanchon/pdfium-binaries/releases/latest/download/pdfium-linux-arm64.tgz" && \ # tar -xzf pdfium-linux-arm64.tgz -C lib --strip-components=1 && \ # rm pdfium-linux-arm64.tgz # Verify the library is correctly placed ls -la lib/libpdfium.so # Build and run the project cargo run

Launch the application - no arguments required
Open a PDF file:
- Click the "📁 PDFを開く" (Open PDF) button in the header
- Select a PDF file from the file dialog
View PDF content:
- All pages are displayed in a continuous scroll format
- Loading progress is shown at the top during initial page rendering
- Individual page rotation controls (90° increments) with persistence
Reading Progress Management:
- Bookmarks: Use "🔖 ブックマーク" to set page-level reading bookmarks
- Position Markers: Enable "📍 マーカーモード" to place precise location markers within pages
- Automatic Navigation: Click markers in "📋 マーカー一覧" for automatic page scrolling
- Recent Files: Access recently opened files via "📋 最近のファイル"
AI Search functionality:
- Select AI provider: Gemini, ChatGPT, or Claude
- Enter your API key once - it will be automatically saved and restored in future sessions
- Type your search query to ask about terms or concepts
- Click "検索" (Search) to get AI-powered explanations
- Searched terms are automatically saved to your flashcard collection
File management:
- Use "❌ 閉じる" (Close) button to close the current PDF
- Open different files without restarting the application

Header Bar: File control buttons and application title
- "🔖 ブックマーク" (Bookmarks) button for reading progress management
- "📍 マーカーモード" (Marker Mode) toggle for position marking
- "📋 マーカー一覧" (Marker List) for viewing and navigating to saved markers
- "📋 最近のファイル" (Recent Files) for quick file access
- "📁 PDFを開く" (Open PDF) button for file selection
- "❌ 閉じる" (Close) button to close current PDF
Status Bar: Loading progress and PDF information (when PDF is loaded)
Main Content Area:
- Left Panel: PDF display area optimized for vertical documents with continuous scroll
- Right Panel: AI search interface
  - AI provider selection (Gemini, ChatGPT, Claude)
  - API key input (password field) with automatic save/restore
  - Search query input for term explanations
  - Search button with loading indicator
  - Results area with clean text formatting
  - Flashcard button to view saved search terms
Welcome Screen: Displayed when no PDF is loaded with usage instructions

The application automatically saves search terms and their AI-generated definitions as flashcards for future reference.

Storage Location:

macOS: ~/.config/pdf-viewer/
Linux: ~/.config/pdf-viewer/
Windows: %USERPROFILE%\.config\pdf-viewer\

Data Files:

flashcards.json - Saved search terms and definitions
recent_files.json - Recently opened PDF files (max 10)
api_keys.json - Saved API keys for AI providers
bookmarks.json - Reading bookmarks for each PDF file
position_markers.json - Precise position markers within PDF pages
page_rotations.json - Individual page rotation states

Data Format: Flashcards are stored in JSON format with the following structure:

[ { "id": "1734493200000-machine-lea", "term": "machine learning", "definition": "AI-generated explanation of the term...", "created_at": "2024-12-18 12:00:00 UTC" } ]

Features:

Automatic saving when AI search is performed
Duplicate term detection (updates existing definitions)
Persistent storage across application sessions
Popup interface for browsing saved flashcards
Individual flashcard deletion capability

Implementation Highlights

PDF pages rendered at 1000x1400 pixels for balanced quality and performance
Continuous scroll display for all pages at once
Automatic color space conversion (BGRA → RGBA)
PNG encoding with base64 data URLs for web display

Asynchronous Gemini API calls that don't block the UI
Markdown formatting removal for clean text display
Real-time search status indicators
Comprehensive error handling

Performance Optimizations

Smart page-level caching system using HashMap with duplicate loading prevention
Optimized parallel page rendering with CPU-aware batch processing
Memory-efficient image handling with color space optimization
Responsive UI updates with non-blocking async operations
Single-load policy to prevent unnecessary re-rendering of same PDF files

Modern Dioxus Architecture

Reactive state management with use_signal
Memoized computations with use_memo
Component-based design patterns
CSS-in-Rust styling approach

├── src/ │ └── main.rs # Main Dioxus application ├── assets/ │ └── style.css # UI styling ├── lib/ │ └── libpdfium.dylib # PDFium library for macOS ├── Cargo.toml # Dependencies and configuration └── README.md # This file

PDF Rendering: render_pdf_page_with_text() function handles high-resolution page rendering with text extraction
File Management: Dynamic PDF loading/closing with native file dialog integration and recent files history
Reading Progress: Bookmark and position marker systems with JSON persistence and automatic navigation
Page Rotation: Individual page rotation controls with state persistence
AI Search: search_with_ai() async function for multi-provider API integration
Text Processing: clean_markdown_text() for formatting search results
UI Components: Reactive Dioxus components with modern state management and responsive layout
JavaScript Integration: eval() function for automatic page scrolling and DOM manipulation