python , Weaviate example

add.py

import weaviate
from weaviate.connect import ConnectionParams
from weaviate.classes.config import Property, DataType
import weaviate.classes.config as Configure

COLLECT_NAME = "document"
client = weaviate.WeaviateClient(
    connection_params=ConnectionParams.from_url(
        "http://localhost:8080",
        grpc_port=50051
    )
)

client.connect()
# 接続確認
print(client.is_ready())

# 1. コレクション（テーブルのようなもの）の作成
# すでに存在する場合は取得、ない場合は作成
if not client.collections.exists(COLLECT_NAME):
    print("not-exist-COLLECT")
else:
    print("exist-COLLECT=" + COLLECT_NAME)
    articles = client.collections.get(COLLECT_NAME)

    articles.data.insert(
    properties={
        "content": "embedding を自分で作る-1",
        "category": "none"
    },
    vector=[0.12, 0.34, 0.41]  # embedding
)


# 処理が終わったら閉じる（または context manager を使用）
client.close()

init.py

import weaviate
from weaviate.connect import ConnectionParams
from weaviate.classes.config import Property, DataType

COLLECT_NAME = "document"
client = weaviate.WeaviateClient(
    connection_params=ConnectionParams.from_url(
        "http://localhost:8080",
        grpc_port=50051
    )
)

client.connect()
# 接続確認
print(client.is_ready())

client.collections.create(
    name=COLLECT_NAME,
    properties=[
        Property(name="content", data_type=DataType.TEXT),
        Property(name="category", data_type=DataType.TEXT),
    ]
)

# 処理が終わったら閉じる（または context manager を使用）
client.close()

search.py

import weaviate
from weaviate.connect import ConnectionParams
from weaviate.classes.config import Property, DataType
from weaviate.classes.query import MetadataQuery
import weaviate.classes.config as Configure

    
COLLECT_NAME = "document"
client = weaviate.WeaviateClient(
    connection_params=ConnectionParams.from_url(
        "http://localhost:8080",
        grpc_port=50051
    )
)

client.connect()
# 接続確認
print(client.is_ready())
try:
    # 1. コレクション（テーブルのようなもの）の作成
    # すでに存在する場合は取得、ない場合は作成
    if not client.collections.exists(COLLECT_NAME):
        print("not-exist-COLLECT")
    else:
        print("exist-COLLECT=" + COLLECT_NAME)
        collection = client.collections.get(COLLECT_NAME)
        query_vector = [0.12, 0.34, 0.36] 
        # 処理が終わったら閉じる（または context manager を使用）
        # 2. 指定したベクトルで近傍検索を実行
        response = collection.query.near_vector(
            near_vector=query_vector,
            limit=2,                             # 取得件数
            return_metadata=MetadataQuery(distance=True) # 距離（近さ）も取得する場合
        )
        # 3. 結果の表示
        for obj in response.objects:
            print(f"content: {obj.properties.get('content')}")
            print(f"距離: {obj.metadata.distance}") # 値が小さいほど似ている
            print("-" * 20)

        client.close()
finally:
    client.close()