milvus_lite_vector_collection_insert_and_similarity_search.py

1from pymilvus import MilvusClient
2import numpy as np
3
4# 1. Set up a local Milvus Lite instance
5# This will create a local file named "milvus_demo.db"
6client = MilvusClient("milvus_demo.db")
7
8# 2. Create a collection in quick setup mode
9# The collection will have a primary field named "id" and a vector field named "vector"
10if client.has_collection(collection_name="demo_collection"):
11    client.drop_collection(collection_name="demo_collection")
12
13client.create_collection(
14    collection_name="demo_collection",
15    dimension=5 # The vectors we will use have 5 dimensions
16)
17
18# 3. Insert data
19# Data is represented as a list of dictionaries
20data = [
21    {"id": 0, "vector": [0.1, 0.2, 0.3, 0.4, 0.5], "color": "pink", "tag": 1},
22    {"id": 1, "vector": [0.2, 0.3, 0.4, 0.5, 0.6], "color": "red", "tag": 2},
23    {"id": 2, "vector": [0.3, 0.4, 0.5, 0.6, 0.7], "color": "blue", "tag": 3},
24]
25
26res = client.insert(
27    collection_name="demo_collection",
28    data=data
29)
30
31print(f"Insert result: {res}")
32
33# 4. Search for similar vectors
34query_vectors = [[0.15, 0.25, 0.35, 0.45, 0.55]]
35
36res = client.search(
37    collection_name="demo_collection",
38    data=query_vectors,
39    limit=2,  # Return top 2 results
40    output_fields=["color"], # Return the "color" field in results
41)
42
43print("Search results:")
44for result in res:
45    print(result)
46
47# 5. Query data with a filter
48res = client.query(
49    collection_name="demo_collection",
50    filter="id > 0",
51    output_fields=["color", "tag"]
52)
53
54print(f"Query result: {res}")
55
56# 6. Delete data
57client.delete(
58    collection_name="demo_collection",
59    filter="id == 0"
60)
61
62# 7. Drop the collection
63client.drop_collection(collection_name="demo_collection")