# ADR-004: Content Quality Management
## Status
Accepted
## Context
The D&D 5e wiki contains mixed content quality with three distinct categories:
- **Official**: Published WotC content (Player's Handbook, DMG, etc.)
- **Unearthed Arcana**: Playtest content from WotC
- **Homebrew**: Community-created content
Users need to understand content source and quality for character building and campaign decisions. The guide analysis shows content can be categorized through text analysis and metadata extraction.
## Decision
Implement a content quality classification system that automatically categorizes content and provides quality metadata to users.
### Quality Categories
1. **Published/Official** - Core rulebooks and official supplements
2. **Unearthed Arcana** - Official playtest content
3. **Homebrew** - Community content
4. **Unknown** - Content that cannot be reliably categorized
### Detection Strategy
```python
def get_content_metadata(self, endpoint: str) -> Dict:
soup = self._fetch_page(endpoint)
text = soup.get_text().lower()
metadata = {
'source': 'Unknown',
'quality': 'Unknown',
'verified': False,
'book_references': []
}
# Detection patterns
if 'unearthed arcana' in text:
metadata['source'] = 'Unearthed Arcana'
metadata['quality'] = 'Playtest'
elif 'homebrew' in text:
metadata['source'] = 'Homebrew'
metadata['quality'] = 'Community'
elif any(book in text for book in OFFICIAL_BOOKS):
metadata['source'] = 'Official'
metadata['quality'] = 'Published'
metadata['verified'] = True
return metadata
```
### Official Source References
Maintain list of official D&D 5e sources:
- Player's Handbook (PHB)
- Dungeon Master's Guide (DMG)
- Monster Manual (MM)
- Xanathar's Guide to Everything (XGtE)
- Tasha's Cauldron of Everything (TCE)
- Volo's Guide to Monsters (VGtM)
- Mordenkainen's Tome of Foes (MTF)
- And other official supplements
### Quality Indicators in API Responses
Include quality metadata in all content responses:
```json
{
"name": "Fireball",
"level": 3,
"school": "Evocation",
"quality": {
"source": "Official",
"verified": true,
"book_references": ["Player's Handbook"],
"reliability": "high"
}
}
```
### User Experience Features
- **Quality Badges**: Visual indicators in responses
- **Filtering Options**: Allow users to filter by content quality
- **Warning System**: Clear warnings for homebrew/playtest content
- **Quality Scores**: Numeric reliability ratings (1-5 scale)
## Alternatives Considered
1. **Manual Curation**: Too labor-intensive and not scalable
2. **Community Voting**: Complex to implement, subjective quality measures
3. **No Quality Tracking**: Leaves users unable to distinguish official from homebrew
4. **External Quality Database**: Would require maintaining separate dataset
## Implementation Details
### Quality Detection Rules
- **Official Content**: Contains references to official books, no homebrew indicators
- **UA Content**: Explicitly labeled "Unearthed Arcana" or links to UA articles
- **Homebrew Content**: Contains "homebrew" text, community attribution, or non-official formatting
- **Uncertain Content**: Missing clear indicators, mixed signals
### Confidence Scoring
- **High Confidence (5)**: Clear official book references, verified formatting
- **Medium Confidence (3)**: Some indicators but not definitive
- **Low Confidence (1)**: Conflicting or minimal quality signals
### Caching Quality Metadata
- Cache quality assessments to avoid repeated analysis
- Update quality cache less frequently than content cache
- Allow manual quality overrides for edge cases
## Consequences
### Positive
- Users can make informed decisions about content reliability
- Enables content filtering for official-only campaigns
- Builds trust through transparency about content sources
- Supports DMs in managing campaign content standards
### Negative
- Additional processing overhead for quality analysis
- Risk of misclassification affecting user experience
- Maintenance burden for official source list updates
- Potential bias against community content
### Quality Risks
- False positives: Homebrew content marked as official
- False negatives: Official content marked as homebrew
- Evolution of content quality over time
- Subjective nature of "quality" assessment
## Monitoring and Maintenance
- Track quality classification accuracy through user feedback
- Regular audits of high-impact content (spells, races, classes)
- Update official source list with new releases
- Monitor for quality classification edge cases
## User Guidelines
- Clearly document quality categories and their meanings
- Provide guidance on using quality filters effectively
- Explain limitations of automated quality detection
- Encourage user verification for campaign-critical content
## Future Enhancements
- Machine learning for improved quality classification
- Community quality ratings and feedback
- Integration with official D&D Beyond API when available
- Quality-based content recommendations
