qdrant_client_collection_creation_and_vector_search_quickstart.py

1from qdrant_client import QdrantClient
2from qdrant_client.models import VectorParams, Distance
3import numpy as np
4
5# Initialize the client
6# Use ":memory:" for a temporary in-memory database
7client = QdrantClient(host="localhost", port=6333)
8
9# 1. Create a collection
10client.recreate_collection(
11    collection_name="test_collection",
12    vectors_config=VectorParams(size=4, distance=Distance.DOT),
13)
14
15# 2. Upsert vectors (data)
16vectors = np.random.rand(100, 4).tolist()
17client.upsert(
18    collection_name="test_collection",
19    points=[
20        {
21            "id": i,
22            "vector": vectors[i],
23            "payload": {"color": "red" if i % 2 == 0 else "green"}
24        }
25        for i in range(100)
26    ],
27)
28
29# 3. Search for similar vectors
30search_result = client.search(
31    collection_name="test_collection",
32    query_vector=[0.2, 0.1, 0.9, 0.7],
33    limit=3
34)
35
36for res in search_result:
37    print(f"ID: {res.id}, Score: {res.score}, Payload: {res.payload}")