qdrant_client_in_memory_collection_and_vector_search.py

1from qdrant_client import QdrantClient
2from qdrant_client.models import VectorParams, Distance, PointStruct
3
4# Initialize the client
5# :memory: allows running Qdrant in-memory without a server
6client = QdrantClient(":memory:")
7
8# Create a collection
9client.create_collection(
10    collection_name="test_collection",
11    vectors_config=VectorParams(size=4, distance=Distance.DOT),
12)
13
14# Add data points
15operation_info = client.upsert(
16    collection_name="test_collection",
17    wait=True,
18    points=[
19        PointStruct(id=1, vector=[0.05, 0.61, 0.76, 0.74], payload={"city": "Berlin"}),
20        PointStruct(id=2, vector=[0.19, 0.81, 0.75, 0.11], payload={"city": "London"}),
21        PointStruct(id=3, vector=[0.36, 0.55, 0.86, 0.40], payload={"city": "Moscow"}),
22        PointStruct(id=4, vector=[0.18, 0.01, 0.85, 0.80], payload={"city": "New York"}),
23        PointStruct(id=5, vector=[0.24, 0.18, 0.22, 0.44], payload={"city": "Tokyo"}),
24    ],
25)
26
27# Search for the most similar vectors
28search_result = client.search(
29    collection_name="test_collection",
30    query_vector=[0.2, 0.1, 0.9, 0.7],
31    limit=3
32)
33
34for result in search_result:
35    print(f"ID: {result.id}, Score: {result.score}, Payload: {result.payload}")