Open Census MCP Server

#!/usr/bin/env python3 """Find all FAISS files and check their dimensions + timestamps""" import os import time from pathlib import Path import faiss def find_all_faiss_files(): """Find all .faiss files in the project and check their dimensions""" print("🔍 Searching for ALL .faiss files") print("=" * 40) project_root = Path("/Users/brock/Documents/GitHub/census-mcp-server") faiss_files = [] # Search recursively for .faiss files for faiss_file in project_root.rglob("*.faiss"): faiss_files.append(faiss_file) if not faiss_files: print("❌ No .faiss files found in project") return print(f"Found {len(faiss_files)} FAISS files:\n") current_time = time.time() for faiss_file in sorted(faiss_files): try: # Get file info stat = faiss_file.stat() size_mb = stat.st_size / (1024 * 1024) modified_time = stat.st_mtime age_hours = (current_time - modified_time) / 3600 # Load FAISS index to check dimensions index = faiss.read_index(str(faiss_file)) dimensions = index.d count = index.ntotal # Determine embedding type if dimensions == 3072: embed_type = "OpenAI ✅" elif dimensions == 768: embed_type = "SentenceTransformers ❌" else: embed_type = f"Unknown ({dimensions})" # Show age if age_hours < 1: age_str = f"{age_hours * 60:.0f}m ago 🔥" elif age_hours < 24: age_str = f"{age_hours:.1f}h ago" else: age_str = f"{age_hours / 24:.1f}d ago" print(f"📁 {faiss_file.relative_to(project_root)}") print(f" Dimensions: {dimensions} ({embed_type})") print(f" Vectors: {count:,}") print(f" Size: {size_mb:.1f} MB") print(f" Modified: {age_str}") print() except Exception as e: print(f"❌ Error reading {faiss_file}: {e}") print() def check_expected_locations(): """Check where kb_search.py expects to find files""" print("🎯 Expected File Locations") print("=" * 30) expected_locations = [ "knowledge-base/variables-faiss/variables.faiss", "knowledge-base/table-catalog/table_embeddings_enhanced.faiss", "knowledge-base/table-catalog/table_embeddings.faiss" ] project_root = Path("/Users/brock/Documents/GitHub/census-mcp-server") for location in expected_locations: full_path = project_root / location if full_path.exists(): try: index = faiss.read_index(str(full_path)) dimensions = index.d embed_type = "OpenAI ✅" if dimensions == 3072 else "SentenceTransformers ❌" print(f"✅ {location}: {dimensions} dims ({embed_type})") except Exception as e: print(f"❌ {location}: Error reading - {e}") else: print(f"❌ {location}: Not found") def find_recent_build_outputs(): """Find any files created in the last few hours""" print(f"\n⏰ Recent Build Outputs (last 2 hours)") print("=" * 40) project_root = Path("/Users/brock/Documents/GitHub/census-mcp-server") current_time = time.time() recent_files = [] # Look for recent files for file_path in project_root.rglob("*"): if file_path.is_file(): try: stat = file_path.stat() age_hours = (current_time - stat.st_mtime) / 3600 # Show files created/modified in last 2 hours if age_hours < 2: size_mb = stat.st_size / (1024 * 1024) recent_files.append((file_path, age_hours, size_mb)) except: continue if recent_files: recent_files.sort(key=lambda x: x[1]) # Sort by age for file_path, age_hours, size_mb in recent_files: age_str = f"{age_hours * 60:.0f}m ago" print(f"📄 {file_path.relative_to(project_root)}") print(f" Size: {size_mb:.1f} MB, Modified: {age_str}") # Check if it's a FAISS file if file_path.suffix == '.faiss': try: index = faiss.read_index(str(file_path)) dimensions = index.d embed_type = "OpenAI ✅" if dimensions == 3072 else "SentenceTransformers ❌" print(f" 🎯 FAISS: {dimensions} dims ({embed_type})") except: pass print() else: print("No files modified in the last 2 hours") if __name__ == "__main__": find_all_faiss_files() check_expected_locations() find_recent_build_outputs()