# Fix for KeyError: 'method' in Risk Discovery Comparison

## Problem
When running `compare_risk_discovery.py`, the script failed with:
```
KeyError: 'method'
```

This occurred because the K-Means implementation (`UnsupervisedRiskDiscovery`) was returning inconsistent data format compared to other methods.

## Root Cause
Different discovery methods were returning different data structures:

### Other Methods (LDA, NMF, etc.) returned:
```python
{
    'method': 'LDA_Topic_Modeling',
    'n_topics': 7,
    'discovered_topics': {...},
    'quality_metrics': {...}
}
```

### K-Means returned:
```python
{
    # Just the patterns dictionary, no metadata
    'pattern_1': {...},
    'pattern_2': {...}
}
```

The comparison function expected all methods to return a consistent structure with metadata.

## Solution

### 1. Fixed K-Means Return Format (`risk_discovery.py`)

**Before:**
```python
def discover_risk_patterns(self, clause_texts: List[str]) -> Dict[str, Any]:
    # ... clustering logic ...
    return self.discovered_patterns  # Just patterns dict
```

**After:**
```python
def discover_risk_patterns(self, clause_texts: List[str]) -> Dict[str, Any]:
    # ... clustering logic ...
    
    # Calculate quality metrics
    from sklearn.metrics import silhouette_score
    try:
        silhouette = silhouette_score(self.feature_matrix, self.cluster_labels)
    except:
        silhouette = 0.0
    
    # Return structured results for comparison
    return {
        'method': 'K-Means_Clustering',
        'n_clusters': self.n_clusters,
        'discovered_patterns': self.discovered_patterns,
        'cluster_labels': self.cluster_labels,
        'quality_metrics': {
            'silhouette_score': silhouette,
            'n_patterns': len(self.discovered_patterns)
        }
    }
```

### 2. Fixed Report Pattern Display (`compare_risk_discovery.py`)

Updated pattern display code to handle different attribute names:

**Before:**
```python
elif 'discovered_patterns' in res:
    report.append("\nTop 3 Patterns:")
    for i, (pattern_id, pattern) in enumerate(list(res['discovered_patterns'].items())[:3]):
        report.append(f"  Pattern {pattern_id}: {pattern.get('name', 'Unnamed')}")
        report.append(f"    Keywords: {', '.join(pattern.get('top_keywords', [])[:5])}")
        report.append(f"    Clauses: {pattern.get('size', 0)}")
```

**After:**
```python
elif 'discovered_patterns' in res:
    report.append("\nTop 3 Patterns:")
    for i, (pattern_id, pattern) in enumerate(list(res['discovered_patterns'].items())[:3]):
        # Handle different pattern formats
        pattern_name = pattern_id if isinstance(pattern_id, str) else pattern.get('name', f'Pattern {pattern_id}')
        keywords = pattern.get('key_terms', pattern.get('top_keywords', []))
        clause_count = pattern.get('clause_count', pattern.get('size', 0))
        
        report.append(f"  {pattern_name}")
        if keywords:
            report.append(f"    Keywords: {', '.join(keywords[:5])}")
        report.append(f"    Clauses: {clause_count}")
```

## Result

All discovery methods now return consistent data structures:

```python
{
    'method': '<method_name>',           # Method identifier
    'n_clusters' or 'n_topics': int,     # Number of patterns
    'discovered_*': {...},                # Pattern details
    'quality_metrics': {...}              # Performance metrics
}
```

## Files Modified
1. `risk_discovery.py` - Updated `discover_risk_patterns()` return value
2. `compare_risk_discovery.py` - Updated pattern display to handle different formats

## Testing
Once dependencies are installed:
```bash
cd /home/deepu/Downloads/code2
pip install -r requirements.txt
python3 compare_risk_discovery.py              # Basic comparison (4 methods)
python3 compare_risk_discovery.py --advanced   # Full comparison (9 methods)
```

## Additional Fixes in This Session
1. **NMF Parameter Compatibility** - Added version detection for scikit-learn API differences
2. **Full Dataset Support** - Removed clause limits, added `--max-clauses` CLI option
3. **Consistent Return Formats** - Standardized all discovery methods

All 9 risk discovery methods should now work correctly!