nniback / core /analytics /use_case_clustering.py
Kyrlos11's picture
Upload 842 files
22087d0 verified
Raw History Blame Contribute Delete
3 kB
"""
Use Case Clustering Service
Groups use cases by category - the LLM already categorizes during analysis,
so we just need a simple GROUP BY query.
"""
from typing import List, Dict, Any, Optional
from core.services.supabase import DBConnection
from core.utils.logger import logger
async def get_clustered_use_cases(
date_from: Optional[str] = None,
date_to: Optional[str] = None,
distance_threshold: float = 0.3, # Kept for API compatibility, not used
min_cluster_size: int = 1
) -> List[Dict[str, Any]]:
"""
Get use cases grouped by category.
The LLM already assigns use_case_category during analysis,
so we just GROUP BY that field and count unique threads.
"""
try:
db = DBConnection()
client = await db.client
# Simple query - group by category, count unique threads
# Filter out both NULL and empty strings
query = client.from_('conversation_analytics')\
.select('use_case_category, thread_id, account_id')\
.not_.is_('use_case_category', None)\
.neq('use_case_category', '')
if date_from:
query = query.gte('analyzed_at', f"{date_from}T00:00:00Z")
if date_to:
query = query.lte('analyzed_at', f"{date_to}T23:59:59Z")
result = await query.execute()
records = result.data or []
# Debug: log actual category values
categories_found = [r.get('use_case_category') for r in records]
logger.info(f"[CLUSTERING] Categories in records: {categories_found}")
# Group by category
from collections import defaultdict
groups = defaultdict(lambda: {'threads': set(), 'examples': []})
for r in records:
cat = r.get('use_case_category')
if cat:
groups[cat]['threads'].add(r.get('thread_id'))
if len(groups[cat]['examples']) < 5:
groups[cat]['examples'].append({
'thread_id': r.get('thread_id'),
'account_id': r.get('account_id')
})
# Debug: log group counts before filtering
logger.info(f"[CLUSTERING] Groups before min_size filter: {[(k, len(v['threads'])) for k, v in groups.items()]}")
# Build result
clusters = []
for cat, data in groups.items():
count = len(data['threads'])
if count >= min_cluster_size:
clusters.append({
'cluster_id': hash(cat) % 10000,
'label': cat,
'count': count,
'examples': data['examples']
})
clusters.sort(key=lambda x: x['count'], reverse=True)
logger.info(f"[CLUSTERING] Found {len(clusters)} categories from {len(records)} records")
return clusters
except Exception as e:
logger.error(f"[CLUSTERING] Failed to get use case clusters: {e}", exc_info=True)
raise