【実装】Azure AI Searchでベクトル検索対応インデックスを作成・移行する方法

from azure.search.documents.indexes import SearchIndexClient
from azure.core.credentials import AzureKeyCredential

client = SearchIndexClient(
    endpoint="[ENDPOINT]",
    credential=AzureKeyCredential("[API_KEY]")
)

# 削除対象のインデックス名を明示的にリストアップ
target_indexes = [
    "myapp_basic_knowledge_20260131",
    "myapp_requirements_definition_20260131",
    "myapp_specification_20260131",
    "myapp_detailed_design_20260131",
]

for name in target_indexes:
    client.delete_index(name)
    print(f"Deleted: {name}")

from azure.search.documents.indexes.models import (
    SearchIndex,
    SearchField,
    SearchFieldDataType,
    VectorSearch,
    HnswAlgorithmConfiguration,
    VectorSearchProfile,
    AzureOpenAIVectorizer,
    AzureOpenAIVectorizerParameters,
)

def create_index_with_vectorizer(client, index_name: str):
    fields = [
        SearchField(name="id", type=SearchFieldDataType.String, key=True),
        SearchField(name="content", type=SearchFieldDataType.String, searchable=True),
        SearchField(name="file_name", type=SearchFieldDataType.String, filterable=True),
        SearchField(name="page", type=SearchFieldDataType.String),
        SearchField(name="url", type=SearchFieldDataType.String),
        SearchField(
            name="content_vector",
            type=SearchFieldDataType.Collection(SearchFieldDataType.Single),
            searchable=True,
            vector_search_dimensions=1536,
            vector_search_profile_name="my-vector-profile",
        ),
    ]

    vector_search = VectorSearch(
        algorithms=[HnswAlgorithmConfiguration(name="my-hnsw")],
        profiles=[
            VectorSearchProfile(
                name="my-vector-profile",
                algorithm_configuration_name="my-hnsw",
                vectorizer_name="my-vectorizer",
            )
        ],
        vectorizers=[
            AzureOpenAIVectorizer(
                vectorizer_name="my-vectorizer",
                parameters=AzureOpenAIVectorizerParameters(
                    resource_url="[AOAI_ENDPOINT]",
                    deployment_name="text-embedding-3-small",
                    model_name="text-embedding-3-small",
                    api_key="[AOAI_API_KEY]",
                ),
            )
        ],
    )

    index = SearchIndex(name=index_name, fields=fields, vector_search=vector_search)
    client.create_or_update_index(index)
    print(f"Created: {index_name}")

LAYER_TO_FOLDER = {
    "requirements": "_requirements_definition_",
    "specifications": "_specification_",
    "detailed_design": "_detailed_design_",
    "explanation": "_basic_knowledge_",
}

for layer, folder in LAYER_TO_FOLDER.items():
    index_name = f"{prefix}{folder}{date_suffix}"
    create_index_with_vectorizer(client, index_name)
    
    # レイヤーに属するチャンクを抽出して登録
    chunks = [c for c in all_chunks if c["layer"] == layer]
    search_client = client.get_search_client(index_name)
    search_client.upload_documents(chunks)

from azure.search.documents.models import VectorizedQuery

def search_with_hybrid(index_name: str, query: str, top: int = 5):
    vector_query = VectorizedQuery(
        vector=generate_embeddings(query),
        k_nearest_neighbors=top,
        fields="content_vector",
    )
    search_client = client.get_search_client(index_name)
    results = search_client.search(
        search_text=query,           # キーワード検索
        vector_queries=[vector_query], # ベクトル検索
        select=["content", "file_name", "page", "url"],
        top=top,
    )
    return results