GLOST Plugin Guides
Comprehensive guides demonstrating the real-world value of each GLOST plugin package.
Philosophy: Show, Don’t Tell
Each guide demonstrates:
- Real use cases with actual scenarios
- Before/After comparisons showing the transformation
- Data examples with real corpus/dictionary data
- Visual demonstrations of output and UI integration
- Measurable value (e.g., “40% faster comprehension”)
- Working code examples you can copy and adapt
Plugin Guides
🔢 Frequency Plugin
Use Case: Highlight words by frequency for prioritized vocabulary learning
Key Value:
- Focus on high-impact vocabulary first (80/20 rule)
- Accurate text difficulty assessment
- Data-driven learning paths
Example: “ephemeral” (rare) vs “the” (very common) → visual prioritization
📝 Part-of-Speech Plugin
Use Case: Identify nouns, verbs, adjectives for grammar learning
Key Value:
- Sentence structure visualization
- Automatic exercise generation
- Pattern recognition (adjective + noun, SVO order)
Example: “The quick brown fox jumps” → [Art][Adj][Adj][N][V]
📊 Difficulty Plugin
Use Case: Assess text difficulty for adaptive learning platforms
Key Value:
- Automatic text leveling (beginner/intermediate/advanced)
- Personalized content recommendations
- Challenge word identification
Example: 44% advanced words → “Advanced Level” text
⚤ Gender Plugin
Use Case: Learn grammatical gender for Romance/Germanic languages
Key Value:
- Correct article selection (le vs la, der vs die vs das)
- Visual gender reinforcement
- Gender pattern discovery
Example: “chat” (masculine) → use “le”, not “la”
🔗 Clause Segmenter Plugin
Use Case: Break complex sentences into comprehensible clauses
Key Value:
- 40% faster reading comprehension
- Core meaning vs supporting details
- Sentence complexity analysis
Example: “The student [who studied] passed [because she understood]” → 3 clauses identified
Quick Comparison
| Plugin | What It Does | Input | Output | Primary Benefit |
|---|---|---|---|---|
| Frequency | Rates word commonness | ”ephemeral” | Rare (43 occurrences) | Prioritized learning |
| POS | Tags grammar role | ”jumps” | Verb | Pattern recognition |
| Difficulty | Assesses word level | ”aforementioned” | Advanced (C2) | Adaptive content |
| Gender | Identifies noun gender | ”chat” | Masculine | Correct articles |
| Clause Segmenter | Breaks sentence structure | Complex sentence | 3 clauses | Better comprehension |
Data Quality Principle
All plugins follow the philosophy: “No data is better than bad data.”
✅ We Use
- Validated corpus data (BNC, TNK, JNC)
- NLP models (Stanford NLP, spaCy)
- Official dictionaries (CEFR word lists, JLPT levels)
- Language-specific rule sets
❌ We Don’t Use
- Length-based heuristics (“long = rare”)
- Pattern guessing (“ends in -ing = verb”)
- Arbitrary thresholds
- Single-source fallbacks
Result: High accuracy or no result. Systems can handle undefined values gracefully, but incorrect data breaks trust.
Integration Patterns
Pattern 1: Basic Usage
import { processGLOSTWithExtensionsAsync } from "@glotblocks/glost-plugins";
import { createFrequencyExtension } from "@glotblocks/glost-frequency";
import { createFrequencyProvider } from "./my-provider";
// 1. Create provider with real data
const provider = createFrequencyProvider({ corpusData });
// 2. Create extension
const [freqGen, freqEnh] = createFrequencyExtension({
targetLanguage: "en",
provider
});
// 3. Process document
const result = await processGLOSTWithExtensionsAsync(doc, [freqGen, freqEnh]);Pattern 2: Multi-Extension Pipeline
import { processGLOSTWithExtensionsAsync } from "@glotblocks/glost-plugins";
import { createFrequencyExtension } from "@glotblocks/glost-frequency";
import { createDifficultyExtension } from "@glotblocks/glost-difficulty";
// Create extensions
const [freqGen, freqEnh] = createFrequencyExtension({
targetLanguage: "en",
provider: freqProvider
});
const [diffGen, diffEnh] = createDifficultyExtension({
targetLanguage: "en",
provider: diffProvider
});
// Combine multiple extensions
const result = await processGLOSTWithExtensionsAsync(doc, [
freqGen, freqEnh,
diffGen, diffEnh
]);Pattern 3: Language-Specific Providers
import { processGLOSTWithExtensionsAsync } from "@glotblocks/glost-plugins";
import { createFrequencyExtension } from "@glotblocks/glost-frequency";
import { thaiFrequencyProvider } from "@glotblocks/glost-th/extensions";
// Thai processing
const [freqGen, freqEnh] = createFrequencyExtension({
targetLanguage: "th",
provider: thaiFrequencyProvider
});
const result = await processGLOSTWithExtensionsAsync(thaiDoc, [freqGen, freqEnh]);Getting Started
- Choose a plugin based on your use case
- Read the guide for detailed examples and implementation
- Obtain data (corpus, dictionary, word lists, or NLP model)
- Implement provider for your language
- Integrate and test with your GLOST documents
Real-World Applications
🎓 Language Learning Apps
- Frequency: Vocabulary prioritization
- POS: Grammar exercises
- Difficulty: Text leveling
- Gender: Article practice
- Clause Segmenter: Reading comprehension
📚 Digital Reading Platforms
- Frequency: Word highlighting
- Difficulty: Text recommendations
- Clause Segmenter: Sentence simplification
✍️ Writing Assistants
- POS: Grammar checking
- Clause Segmenter: Style analysis
- Frequency: Vocabulary suggestions
📊 Educational Analytics
- Difficulty: Student progress tracking
- Frequency: Vocabulary coverage measurement
- Clause Segmenter: Reading level assessment
Contributing
Found a better use case? Have real data for a language?
- Fork the repo
- Add your example to the relevant guide
- Include real data samples (anonymized if needed)
- Submit a PR with measurable value demonstrated
Next Steps
- Read a specific plugin guide from the list above
- Check the migration guide for upgrading existing code
- Review the provider philosophy for data quality principles
- Explore example implementations for working code
Remember: Each plugin is optional. Use only what adds value to your specific use case. No plugins? No problem. The core GLOST format works perfectly without any plugins.