Bappadala Rohith Kumar Naidu
feat: complete enterprise-grade dataset sync of RAG, offline bundles, and pipeline scripts
92cf271
Raw
History Blame Contribute Delete
1.13 kB
"""Sample random RAG chunks for human quality review."""
import os
import random
CHROMA_PATH = os.getenv("CHROMA_PERSIST_DIR", "./data/chroma_db")
COLLECTION_NAME = os.getenv("CHROMA_COLLECTION_NAME", "safevixai_rag")
def audit_chunks():
import chromadb
client = chromadb.PersistentClient(path=CHROMA_PATH)
col = client.get_collection(COLLECTION_NAME)
count = col.count()
print(f"Total chunks: {count}")
all_data = col.get(limit=count)
if not all_data.get("documents"):
print("No documents found")
return
indices = random.sample(range(len(all_data["documents"])), min(5, count))
for i in indices:
doc = all_data["documents"][i]
meta = all_data["metadatas"][i] if all_data.get("metadatas") else {}
print(f"\n{'='*60}")
print(f"CHUNK {i}")
print(f"Source: {meta.get('source_file', 'unknown')}")
print(f"Section: {meta.get('section', 'unknown')}")
print(f"Category: {meta.get('category', 'unknown')}")
print(f"{'='*60}")
print(doc[:300])
print("...")
if __name__ == "__main__":
audit_chunks()