How benchristel/audition works
Main Components to Include:
Input Processing Layer
• Three main input file types:
• .au files (source files with glosses)
• lexicon.csv (word definitions)
• morphology.yaml (inflection rules)
• generator.txt (word generation rules)
Core Processing Components
• Parser System
• PEG.js parsers (gloss-parser, generator-rule-parser, csv-parser)
• Generator System
• Word generation component
• Morphological rules processor
• Translator System
• Lexicon processor
• Inflection engine
Output Processing Layer
• Generated conlang text output
• Updated lexicon with generated words
• [Input Files] → [Processing Layer] → [Output Files]
Main data flows
• [.au files] ───┐
• [lexicon.csv] ─┼→ [Parser Layer] → [Translator] → [Output Generator] → [Final Text]
• [morphology.yaml] ┘
Key Architectural Elements to Highlight:
Parser Components (src/generated/)
• How the PEG.js grammar files are compiled into parsers
• Relationship between source grammars and generated parsers
Core Processing Pipeline
• Modular structure reflected in the src/ directory
• How components like expression.ts, generator.ts, and translator.ts interact
Library Support Layer (src/lib/)
• Utility modules supporting the main processing pipeline
• Utilities like weighted-random-variable.ts for word generation
Testing Infrastructure
• Test directory relationship to main components
• How test-language/ provides example inputs
Use rectangles for main processing components
• File reading/writing operations
• Data transformation flows
• Component dependencies
Important Relationships to Emphasize:
The pipeline nature of the tool
• Parse → Generate → Transform → Output
The bidirectional relationship between the lexicon and generator
• Generator fills in missing translations
• Lexicon provides existing words
The hierarchical relationship between components
• Main Application
• Parser System
• Generator System
• Translator System
• Lexicon Processor
• Morphology Processor
Level of Detail:
• Keep the high-level flow clear while showing the key processing steps
• File formats at interaction points
• Error handling paths
• Where configuration options affect processing