Milvus Python SDK

Using Milvus python sdk for Milvus

Download

Pymilvus only supports python >= 3.4, is fully tested under 3.4, 3.5, 3.6, 3.7.

Pymilvus can be downloaded using pip. If no use, try pip3

$ pip install pymilvus

Upgrade to newest version

$ pip install --upgrade pymilvus

Import

from milvus import Milvus, IndexType, Status

Getting started

Initial a Milvus instance and connect to the sever

>>> milvus = Milvus()

>>> milvus.connect(host='SERVER-HOST', port='SERVER-PORT')
Status(code=0, message="Success")

Once successfully connected, you can get the version of server

>>> milvus.server_version()
(Status(code=0, message='Success'), 0.3.0)  # this is example version, the real version may vary

Add a new table

First set param

>>> param = {'table_name'='test01', 'dimension'=256, 'index_type'=IndexType.FLAT, 'store_raw_vector'=False}

Then create table

>>> milvus.create_table(param)
Status(message='Table test01 created!', code=0)

Describe the table we just created

>>> milvus.describe_table('test01')
(Status(code=0, message='Success!'), TableSchema(table_name='test01',dimension=256, index_type=1, store_raw_vector=False))

Add vectors into table test01

First create 20 vectors of 256-dimension.

Note that random and pprint we used here is for creating fake vectors data and pretty print, you may not need them in your project

>>> import random
>>> from pprint import pprint

>>> dim = 256  # Dimension of the vector

# Initialize 20 vectors of 256-dimension
>>> fake_vectors = [[random.random() for _ in range(dim)] for _ in range(20)]

Then add vectors into table test01

>>> status, ids = milvus.add_vectors(table_name='test01', records=vectors)
>>> print(status)
Status(code=0, message='Success')
>>> pprint(ids) # List of ids returned
23455321135511233
12245748929023489
...

Search vectors

# create 5 vectors of 256-dimension
>>> q_records = [[random.random() for _ in range(dim)] for _ in range(5)]

Then get results

>>> status, results = milvus.search_vectors(table_name='test01', query_records=q_records, top_k=10)
>>> print(status)
Status(code=0, message='Success')
>>> pprint(results) # Searched top_k vectors

Delet the table we just created

>>> milvus.delete_table(table_name='test01')
Status(code=0, message='Success')

Disconnect with the server

>>> milvus.disconnect()
Status(code=0, message='Success')

Example python

There are some small examples in examples/, you can find more guide there.

Build docs

$ sphinx-build -b html doc/en/ doc/en/build

If you encounter any problems or bugs, please add new issues

Release files for pymilvus 0.1.21

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distribution (wheel)

Table of built distributions (wheels) for pymilvus 0.1.21
File	Interpreter	ABI	Platform	Reset
pymilvus-0.1.21-py3-none-any.whl 25.2 kB Jul 9, 2019	Python 3	none	any	Details

Release files / pymilvus-0.1.21-py3-none-any.whl

Download URL	pymilvus-0.1.21-py3-none-any.whl
Size	25.2 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`61dde56b23c403faebe9db91499b57103a34230c50e47cb5bff406a3732bdf41`
BLAKE2b-256 checksum How to use checksums	`ff7f22ff69a060dc06c55d442c0aeef166f134b73e532dc6d1b1abc66f75aa58`
Upload date	Jul 9, 2019
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/1.13.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/41.0.1 requests-toolbelt/0.9.1 tqdm/4.32.2 CPython/3.7.4`