#!/usr/bin/env python3
"""
Find which Qdrant collection contains embeddings for the current codebase.
"""
import sys
from pathlib import Path
from mcp_server.core.path_utils import PathUtils
# Add project root to Python path
sys.path.insert(0, str(Path(__file__).parent.parent))
from qdrant_client import QdrantClient
def find_codebase_embeddings():
"""Find which collection has current codebase embeddings."""
print("=" * 60)
print("SEARCHING FOR CURRENT CODEBASE EMBEDDINGS")
print("=" * 60)
qdrant_path = "PathUtils.get_workspace_root()/.indexes/qdrant/main.qdrant"
# Remove stale lock if it exists
lock_file = Path(qdrant_path) / ".lock"
if lock_file.exists():
try:
lock_file.unlink()
except OSError:
pass
try:
client = QdrantClient(path=qdrant_path)
collections = client.get_collections()
codebase_patterns = [
"mcp_server",
"code-index-mcp",
"code_index_mcp",
"/workspaces/code-index-mcp",
"dispatcher",
"semantic_indexer",
"plugin_factory"
]
print(f"Searching patterns: {codebase_patterns[:3]}...\n")
results = {}
for collection in collections.collections:
try:
info = client.get_collection(collection.name)
if info.points_count == 0:
continue
# Sample more points to get better coverage
sample_size = min(info.points_count, 200)
sample = client.scroll(
collection_name=collection.name,
limit=sample_size,
with_payload=True
)
if not sample[0]:
continue
matches = 0
matched_files = []
for point in sample[0]:
file_path = (
point.payload.get("file") or
point.payload.get("relative_path") or
point.payload.get("filepath") or
point.payload.get("path") or
""
).lower()
# Check all payload keys
for key, value in point.payload.items():
if isinstance(value, str):
value_lower = value.lower()
if any(pattern in value_lower for pattern in codebase_patterns):
matches += 1
if file_path and len(matched_files) < 5:
matched_files.append(file_path)
break
if matches > 0:
match_rate = (matches / sample_size) * 100
results[collection.name] = {
"matches": matches,
"sample_size": sample_size,
"match_rate": match_rate,
"total_points": info.points_count,
"matched_files": matched_files
}
except Exception as e:
print(f"Error checking {collection.name}: {e}")
# Display results
if results:
print("Collections containing current codebase files:\n")
sorted_results = sorted(results.items(), key=lambda x: x[1]["match_rate"], reverse=True)
for collection_name, data in sorted_results:
print(f"Collection: {collection_name}")
print(f" Total points: {data['total_points']}")
print(f" Matches: {data['matches']}/{data['sample_size']} ({data['match_rate']:.1f}%)")
if data['matched_files']:
print(f" Sample matched files:")
for i, file in enumerate(data['matched_files'][:3]):
print(f" {i+1}. {file}")
print()
# Recommend the best collection
best_collection = sorted_results[0][0]
print(f"🎯 RECOMMENDATION: Use collection '{best_collection}'")
print(f" This collection has the highest match rate for current codebase files.")
else:
print("❌ No collections found containing current codebase files!")
print(" All embeddings may be using different file paths or the index is empty.")
# Try to inspect payload structure
print("\nInspecting payload structure of code-embeddings collection...")
try:
sample = client.scroll(
collection_name="code-embeddings",
limit=3,
with_payload=True
)
if sample[0]:
print("Sample payload keys:")
for i, point in enumerate(sample[0][:3]):
print(f"\nPoint {i+1}:")
for key, value in point.payload.items():
print(f" {key}: {value if len(str(value)) < 100 else str(value)[:100] + '...'}")
except Exception as e:
print(f"Could not inspect code-embeddings: {e}")
except Exception as e:
print(f"❌ Failed to connect to Qdrant: {e}")
print("\n" + "=" * 60)
print("SEARCH COMPLETE")
print("=" * 60)
if __name__ == "__main__":
find_codebase_embeddings()
