383201cacd
## v0.9.20260325_144654 ### Features - API Key Authentication System - Job Worker System - V2 Backup Versioning ### Bug Fixes - get_processor_results_by_job column mapping Co-authored-by: OpenCode
134 lines
3.4 KiB
Python
134 lines
3.4 KiB
Python
#!/opt/homebrew/bin/python3.11
|
|
"""
|
|
Vector Search Test with nomic-embed-text:v1.5 using prefixes - with text content
|
|
"""
|
|
|
|
import time
|
|
import requests
|
|
import psycopg2
|
|
|
|
|
|
VIDEO_UUID = "39567a0eb16f39fd"
|
|
|
|
POSTGRES_CONFIG = {
|
|
"host": "localhost",
|
|
"port": 5432,
|
|
"user": "accusys",
|
|
"password": "Test3200",
|
|
"database": "momentry",
|
|
}
|
|
|
|
MODEL = "nomic-embed-text:v1.5"
|
|
QDRANT_COLLECTION = "chunks_v2"
|
|
|
|
|
|
def get_embedding(text, prefix=""):
|
|
prompt = f"{prefix}{text}"
|
|
resp = requests.post(
|
|
"http://localhost:11434/api/embeddings", json={"model": MODEL, "prompt": prompt}
|
|
)
|
|
return resp.json()["embedding"]
|
|
|
|
|
|
def get_text_from_chunk_id(chunk_id):
|
|
"""Get text from PostgreSQL using chunk_id"""
|
|
conn = psycopg2.connect(**POSTGRES_CONFIG)
|
|
cur = conn.cursor()
|
|
cur.execute("SELECT content->>'text' FROM chunks WHERE chunk_id = %s", (chunk_id,))
|
|
result = cur.fetchone()
|
|
cur.close()
|
|
conn.close()
|
|
return result[0] if result else ""
|
|
|
|
|
|
def test_queries(queries, use_prefix=True):
|
|
"""Test queries against Qdrant"""
|
|
prefix = "search_query: " if use_prefix else ""
|
|
|
|
for query in queries:
|
|
embedding = get_embedding(query, prefix)
|
|
|
|
start = time.time()
|
|
resp = requests.post(
|
|
f"http://localhost:6333/collections/{QDRANT_COLLECTION}/points/search",
|
|
headers={"api-key": "Test3200Test3200Test3200"},
|
|
json={"vector": embedding, "limit": 3},
|
|
)
|
|
elapsed = (time.time() - start) * 1000
|
|
|
|
results = resp.json().get("result", [])
|
|
|
|
print(f"\nQuery: '{query}' ({elapsed:.1f}ms)")
|
|
print("-" * 60)
|
|
for i, r in enumerate(results):
|
|
score = r.get("score", 0)
|
|
# Try to get chunk_id from payload
|
|
payload = r.get("payload", {})
|
|
chunk_id = payload.get("chunk_id", "")
|
|
if not chunk_id:
|
|
# Try to get text from Qdrant payload
|
|
text = payload.get("text", "")[:50]
|
|
else:
|
|
# Get text from PostgreSQL
|
|
text = get_text_from_chunk_id(chunk_id)[:50]
|
|
print(f" {i + 1}. [{score:.3f}] {text}...")
|
|
|
|
|
|
# English queries
|
|
ENGLISH_QUERIES = [
|
|
"a person talking",
|
|
"someone speaking on camera",
|
|
"outdoor scene",
|
|
"indoor setting",
|
|
"walking or moving",
|
|
"dialogue or conversation",
|
|
"looking at something",
|
|
"happy or joyful",
|
|
"serious or dramatic",
|
|
"comedy or funny",
|
|
"wearing a tie",
|
|
"holding an object",
|
|
"sitting on a chair",
|
|
"city or urban",
|
|
"building or room",
|
|
"open space",
|
|
]
|
|
|
|
# Chinese queries
|
|
CHINESE_QUERIES = [
|
|
"有人在說話",
|
|
"戶外場景",
|
|
"室內場景",
|
|
"走路或移動",
|
|
"對話或交談",
|
|
"看著某樣東西",
|
|
"快樂或開心",
|
|
"嚴肅或戲劇性",
|
|
"喜劇或有趣",
|
|
"戴著領帶",
|
|
"拿著東西",
|
|
"坐在椅子上",
|
|
"城市或都市",
|
|
"建築物或房間",
|
|
"開放空間",
|
|
]
|
|
|
|
|
|
if __name__ == "__main__":
|
|
print("=" * 70)
|
|
print(f"Testing with {QDRANT_COLLECTION}")
|
|
print(f"Model: {MODEL}")
|
|
print("Prefix for chunks: search_document:")
|
|
print("Prefix for queries: search_query:")
|
|
print("=" * 70)
|
|
|
|
print("\n" + "=" * 70)
|
|
print("ENGLISH QUERIES")
|
|
print("=" * 70)
|
|
test_queries(ENGLISH_QUERIES)
|
|
|
|
print("\n" + "=" * 70)
|
|
print("CHINESE QUERIES")
|
|
print("=" * 70)
|
|
test_queries(CHINESE_QUERIES)
|