How rmzelle/ref-extractor works
Main Components
a. Client-Side Application
• Browser Environment (primary runtime)
• File Input Handler
• Document Parser
• Reference Extractor Core
• Format Converter
• Output Generator
• User Interface Layer
b. External Dependencies
• Citation.js Library
• Zotero Integration
• Mendeley Integration
Data transformation steps
• File Upload Interface
• Document Parser (for .docx and .odt)
• Reference Extraction Engine
• Format Converter
• Output Generator
• UI Components
• Zotero
• Mendeley
• Citation.js
• Document → Parser
• Parser → Reference Extractor
• Reference Extractor → Format Converter
• Format Converter → Output Generator
D. Output Formats (use parallelograms):
• CSL JSON
• BibTeX
• RIS
• APA Bibliography
• Blue: Core processing components
• Green: Input/Output interfaces
• Orange: External systems
• Grey: Data formats
Important Architectural Considerations:
Client-Side Processing
• That all processing happens locally (privacy feature)
• Clear boundaries of browser environment
Integration Points
• How the system interfaces with Zotero and Mendeley
• Citation.js library integration
Data Flow
• Document the transformation pipeline
• Format conversion paths
Special Notes:
Privacy-First Design
• A note that all processing occurs client-side
• No server communication required
Format Handling
• Supported input formats (.docx, .odt)
• Possible output formats
Browser Integration
• Browser API dependencies
• File system interaction points
Level of Detail:
• Major components
• Data flow directions
• Format transformations
• Integration points
• Security boundaries
Remember to include a notes section explaining:
• Privacy-first architecture
• Browser-based limitations
• Format conversion fidelity
• Integration requirements
This will provide a comprehensive view of the system while maintaining clarity about its browser-based, privacy-focused nature.