Skip to content

Commit

Permalink
Refine interfaces for milvus_client
Browse files Browse the repository at this point in the history
The main focus of this PR is to port some interfaces from ORM
and transform them from an object-oriented form to a functional form.
As these are new interfaces of MilvusClient, certain parameters and return values
have been modified compared to the original ORM interfaces.

Signed-off-by: zhenshan.cao <[email protected]>
  • Loading branch information
czs007 committed Jan 15, 2024
1 parent 5de9116 commit 2f22526
Show file tree
Hide file tree
Showing 18 changed files with 1,165 additions and 283 deletions.
63 changes: 0 additions & 63 deletions examples/hello_milvus_simple.py

This file was deleted.

47 changes: 0 additions & 47 deletions examples/hello_milvus_simple2.py

This file was deleted.

67 changes: 67 additions & 0 deletions examples/milvus_client/alias.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,67 @@
import time
import numpy as np
from pymilvus import (
MilvusClient,
)

fmt = "\n=== {:30} ===\n"
dim = 8
collection_name = "hello_milvus"
milvus_client = MilvusClient("http://localhost:19530")
milvus_client.drop_collection(collection_name)
milvus_client.create_collection(collection_name, dim, consistency_level="Strong", metric_type="L2", auto_id=True)

collection_name2 = "hello_milvus2"
milvus_client.drop_collection(collection_name2)
milvus_client.create_collection(collection_name2, dim, consistency_level="Strong", metric_type="L2", auto_id=True)


print("collections:", milvus_client.list_collections())

desc_c1 = milvus_client.describe_collection(collection_name)
print(f"{collection_name} :", desc_c1)

rng = np.random.default_rng(seed=19530)

rows = [
{"vector": rng.random((1, dim))[0], "a": 100},
{"vector": rng.random((1, dim))[0], "b": 200},
{"vector": rng.random((1, dim))[0], "c": 300},
]

print(fmt.format(f"Start inserting entities to {collection_name}"))
insert_result = milvus_client.insert(collection_name, rows)
print(insert_result)

rows = [
{"vector": rng.random((1, dim))[0], "d": 400},
{"vector": rng.random((1, dim))[0], "e": 500},
{"vector": rng.random((1, dim))[0], "f": 600},
]

print(fmt.format(f"Start inserting entities to {collection_name2}"))
insert_result2 = milvus_client.insert(collection_name2, rows)
print(insert_result2)

alias = "alias_hello_milvus"
milvus_client.create_alias(collection_name, alias)

assert milvus_client.describe_collection(alias) == milvus_client.describe_collection(collection_name)

milvus_client.alter_alias(collection_name2, alias)
assert milvus_client.describe_collection(alias) == milvus_client.describe_collection(collection_name2)

query_results = milvus_client.query(alias, filter= "f == 600")
print("results of query 'f == 600' is ")
for ret in query_results:
print(ret)


milvus_client.drop_alias(alias)
has_collection = milvus_client.has_collection(alias)
assert not has_collection
has_collection = milvus_client.has_collection(collection_name2)
assert has_collection

milvus_client.drop_collection(collection_name)
milvus_client.drop_collection(collection_name2)
70 changes: 70 additions & 0 deletions examples/milvus_client/customize_schema.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,70 @@
import time
import numpy as np
from pymilvus import (
MilvusClient,
DataType
)

fmt = "\n=== {:30} ===\n"
dim = 8
collection_name = "hello_milvus"
milvus_client = MilvusClient("http://localhost:19530")

has_collection = milvus_client.has_collection(collection_name, timeout=5)
if has_collection:
milvus_client.drop_collection(collection_name)

schema = milvus_client.create_schema(enable_dynamic_field=True)
schema.add_field("id", DataType.INT64, is_primary=True)
schema.add_field("embeddings", DataType.FLOAT_VECTOR, dim=dim)
schema.add_field("title", DataType.VARCHAR, max_length=64)


index_params = milvus_client.prepare_index_params()
index_params.add_index(field_name = "embeddings", metric_type="L2")
milvus_client.create_collection(collection_name, schema=schema, index_params=index_params, consistency_level="Strong")

print(fmt.format(" all collections "))
print(milvus_client.list_collections())

print(fmt.format(f"schema of collection {collection_name}"))
print(milvus_client.describe_collection(collection_name))

rng = np.random.default_rng(seed=19530)
rows = [
{"id": 1, "embeddings": rng.random((1, dim))[0], "a": 100, "title": "t1"},
{"id": 2, "embeddings": rng.random((1, dim))[0], "b": 200, "title": "t2"},
{"id": 3, "embeddings": rng.random((1, dim))[0], "c": 300, "title": "t3"},
{"id": 4, "embeddings": rng.random((1, dim))[0], "d": 400, "title": "t4"},
{"id": 5, "embeddings": rng.random((1, dim))[0], "e": 500, "title": "t5"},
{"id": 6, "embeddings": rng.random((1, dim))[0], "f": 600, "title": "t6"},
]

print(fmt.format("Start inserting entities"))
insert_result = milvus_client.insert(collection_name, rows)
print(fmt.format("Inserting entities done"))
print(insert_result)


print(fmt.format("Start load collection "))
milvus_client.load_collection(collection_name)

print(fmt.format("Start query by specifying primary keys"))
query_results = milvus_client.query(collection_name, ids=[2])
print(query_results[0])

print(fmt.format("Start query by specifying filtering expression"))
query_results = milvus_client.query(collection_name, filter= "f == 600 or title == 't2'")
for ret in query_results:
print(ret)

rng = np.random.default_rng(seed=19530)
vectors_to_search = rng.random((1, dim))

print(fmt.format(f"Start search with retrieve serveral fields."))
result = milvus_client.search(collection_name, vectors_to_search, limit=3, output_fields=["pk", "a", "b"])
for hits in result:
for hit in hits:
print(f"hit: {hit}")

milvus_client.drop_collection(collection_name)
87 changes: 87 additions & 0 deletions examples/milvus_client/index.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,87 @@
import time
import numpy as np
from pymilvus import (
MilvusClient,
DataType
)

fmt = "\n=== {:30} ===\n"
dim = 8
collection_name = "hello_milvus"
milvus_client = MilvusClient("http://localhost:19530")

has_collection = milvus_client.has_collection(collection_name, timeout=5)
if has_collection:
milvus_client.drop_collection(collection_name)

schema = milvus_client.create_schema(enable_dynamic_field=True)
schema.add_field("id", DataType.INT64, is_primary=True)
schema.add_field("embeddings", DataType.FLOAT_VECTOR, dim=dim)
schema.add_field("title", DataType.VARCHAR, max_length=64)

# collection is not loaded after creation
milvus_client.create_collection(collection_name, schema=schema, consistency_level="Strong")

rng = np.random.default_rng(seed=19530)
rows = [
{"id": 1, "embeddings": rng.random((1, dim))[0], "a": 100, "title": "t1"},
{"id": 2, "embeddings": rng.random((1, dim))[0], "b": 200, "title": "t2"},
{"id": 3, "embeddings": rng.random((1, dim))[0], "c": 300, "title": "t3"},
{"id": 4, "embeddings": rng.random((1, dim))[0], "d": 400, "title": "t4"},
{"id": 5, "embeddings": rng.random((1, dim))[0], "e": 500, "title": "t5"},
{"id": 6, "embeddings": rng.random((1, dim))[0], "f": 600, "title": "t6"},
]

print(fmt.format("Start inserting entities"))
insert_result = milvus_client.insert(collection_name, rows)
print(fmt.format("Inserting entities done"))
print(insert_result)

index_params = milvus_client.prepare_index_params()
index_params.add_index(field_name = "embeddings", metric_type="L2")
index_params.add_index(field_name = "title", index_type = "TRIE", index_name="my_trie")

print(fmt.format("Start create index"))
milvus_client.create_index(collection_name, index_params)


index_names = milvus_client.list_indexes(collection_name)
print(f"index names for {collection_name}:", index_names)
for index_name in index_names:
index_info = milvus_client.describe_index(collection_name, index_name=index_name)
print(f"index info for index {index_name} is:", index_info)

print(fmt.format("Start load collection"))
milvus_client.load_collection(collection_name)

print(fmt.format("Start query by specifying primary keys"))
query_results = milvus_client.query(collection_name, ids=[2])
print(query_results[0])

print(fmt.format("Start query by specifying filtering expression"))
query_results = milvus_client.query(collection_name, filter= "f == 600 or title == 't2'")
for ret in query_results:
print(ret)

vectors_to_search = rng.random((1, dim))
print(fmt.format(f"Start search with retrieve serveral fields."))
result = milvus_client.search(collection_name, vectors_to_search, limit=3, output_fields=["title"])
for hits in result:
for hit in hits:
print(f"hit: {hit}")



field_index_names = milvus_client.list_indexes(collection_name, field_name = "embeddings")
print(f"index names for {collection_name}`s field embeddings:", field_index_names)

try:
milvus_client.drop_index(collection_name, "my_trie")
except Exception as e:
print(f"cacthed {e}")

milvus_client.release_collection(collection_name)

milvus_client.drop_index(collection_name, "my_trie")

milvus_client.drop_collection(collection_name)
Loading

0 comments on commit 2f22526

Please sign in to comment.