PyMilvus Quick Start

This chapter introduces how to quickly get started with Milvus development using Python.

By running the sample code we provide, you will have a preliminary understanding of Milvus's functions.

Version Requirements

• Milvus 2.3.0
• Python 3 (3.7.1 or higher)
• PyMilvus 2.3.x

Install Milvus Python SDK

python3 -m pip install pymilvus==

Download Sample Code

Use the following command to directly download hello_milvus.py or use the following command for downloading.

$ wget https://www.tizi365.com/storage/hello_milvus.py

Explanation of Sample Code

The sample code performs the following steps.

• Import the PyMilvus package:

from pymilvus import (
    connections,
    utility,
    FieldSchema,
    CollectionSchema,
    DataType,
    Collection,
)

• Connect to the server:

connections.connect("default", host="localhost", port="19530")

• Create a collection:

fields = [
    FieldSchema(name="pk", dtype=DataType.INT64, is_primary=True, auto_id=False),
    FieldSchema(name="random", dtype=DataType.DOUBLE),
    FieldSchema(name="embeddings", dtype=DataType.FLOAT_VECTOR, dim=8)
]
schema = CollectionSchema(fields, "hello_milvus is the simplest demo example, used to introduce the API")
hello_milvus = Collection("hello_milvus", schema)

• Insert vectors into the collection:

import random
entities = [
    [i for i in range(3000)],  # field pk
    [float(random.randrange(-20, -10)) for _ in range(3000)],  # field random
    [[random.random() for _ in range(8)] for _ in range(3000)],  # field embeddings
]
insert_result = hello_milvus.insert(entities)
hello_milvus.flush()

• Create an index on the entities:

index = {
    "index_type": "IVF_FLAT",
    "metric_type": "L2",
    "params": {"nlist": 128},
}
hello_milvus.create_index("embeddings", index)

• Load the collection into memory and perform vector similarity search:

hello_milvus.load()
vectors_to_search = entities[-1][-2:]
search_params = {
    "metric_type": "L2",
    "params": {"nprobe": 10},
}
result = hello_milvus.search(vectors_to_search, "embeddings", search_params, limit=3, output_fields=["random"])

• Perform vector queries:

result = hello_milvus.query(expr="random > -14", output_fields=["random", "embeddings"])

• Perform mixed search:

result = hello_milvus.search(vectors_to_search, "embeddings", search_params, limit=3, expr="random > -12", output_fields=["random"])

• Delete entities based on primary key:

expr = f"pk in [{entities[0][0]}, {entities[0][1]}]"
hello_milvus.delete(expr)

• Delete the collection:

utility.drop_collection("hello_milvus")

Running the sample code

Execute the following command to run the sample code.

$ python3 hello_milvus.py

Below shows the returned results and query latency:

=== Connecting to Milvus ===

Does the collection named hello_milvus exist in Milvus: False

=== Creating collection `hello_milvus` ===

=== Inserting entities ===

Number of entities in Milvus: 3000

=== Creating IVF_FLAT index ===

=== Loading ===

=== Searching based on vector similarity ===

Match: (Distance: 0.0, ID: 2998), Random Field: -11.0
Match: (Distance: 0.11455299705266953, ID: 1581), Random Field: -18.0
Match: (Distance: 0.1232629269361496, ID: 2647), Random Field: -13.0
Match: (Distance: 0.0, ID: 2999), Random Field: -11.0
Match: (Distance: 0.10560893267393112, ID: 2430), Random Field: -18.0
Match: (Distance: 0.13938161730766296, ID: 377), Random Field: -14.0
Search latency = 0.2796 seconds

=== Querying using `random > -14` ===

Query result:
-{'pk': 9, 'random': -13.0, 'embeddings': [0.298433, 0.931987, 0.949756, 0.598713, 0.290125, 0.094323, 0.064444, 0.306993]}
Search latency = 0.2970 seconds

=== Performing mixed search using `random > -12` ===

Match: (Distance: 0.0, ID: 2998), Random Field: -11.0
Match: (Distance: 0.15773043036460876, ID: 472), Random Field: -11.0
Match: (Distance: 0.3273330628871918, ID: 2146), Random Field: -11.0
Match: (Distance: 0.0, ID: 2999), Random Field: -11.0
Match: (Distance: 0.15844076871871948, ID: 2218), Random Field: -11.0
Match: (Distance: 0.1622171700000763, ID: 1403), Random Field: -11.0
Search latency = 0.3028 seconds

=== Deleting using expression `pk in [0, 1]` ===

Querying before deletion based on expression `pk in [0, 1]` -> Result:
-{'pk': 0, 'random': -18.0, 'embeddings': [0.142279, 0.414248, 0.378628, 0.971863, 0.535941, 0.107011, 0.207052, 0.98182]}
-{'pk': 1, 'random': -15.0, 'embeddings': [0.57512, 0.358512, 0.439131, 0.862369, 0.083284, 0.294493, 0.004961, 0.180082]}

Querying after deletion based on expression `pk in [0, 1]` -> Result: []

=== Deleting collection `hello_milvus` ===