Azure Functions 的 Azure OpenAI 语义搜索输入绑定

此示例演示如何对文件执行语义搜索。

[Function("PromptFile")]
public static IActionResult PromptFile(
    [HttpTrigger(AuthorizationLevel.Function, "post")] SemanticSearchRequest unused,
    [SemanticSearchInput("AISearchEndpoint", "openai-index", Query = "{prompt}", ChatModel = "%CHAT_MODEL_DEPLOYMENT_NAME%", EmbeddingsModel = "%EMBEDDING_MODEL_DEPLOYMENT_NAME%")] SemanticSearchContext result)
{
    return new ContentResult { Content = result.Response, ContentType = "text/plain" };
}

此示例演示如何对文件执行语义搜索。

@FunctionName("PromptFile")
public HttpResponseMessage promptFile(
    @HttpTrigger(
        name = "req", 
        methods = {HttpMethod.POST},
        authLevel = AuthorizationLevel.ANONYMOUS)
        HttpRequestMessage<SemanticSearchRequest> request,
    @SemanticSearch(name = "search", searchConnectionName = "AISearchEndpoint", collection = "openai-index", query = "{prompt}", chatModel = "%CHAT_MODEL_DEPLOYMENT_NAME%", embeddingsModel = "%EMBEDDING_MODEL_DEPLOYMENT_NAME%", isReasoningModel = false ) String semanticSearchContext,
    final ExecutionContext context) {
        String response = new JSONObject(semanticSearchContext).getString("Response");
        return request.createResponseBuilder(HttpStatus.OK)
        .header("Content-Type", "application/json")
        .body(response)
        .build();        
}
public class SemanticSearchRequest {
    public String prompt;
    public String getPrompt() {
        return prompt;
    }
    public void setPrompt(String prompt) {
        this.prompt = prompt;
    }        
}

此示例演示如何对文件执行语义搜索。

const semanticSearchInput = input.generic({
    type: "semanticSearch",
    connectionName: "AISearchEndpoint",
    collection: "openai-index",
    query: "{prompt}",
    chatModel: "%CHAT_MODEL_DEPLOYMENT_NAME%",
    embeddingsModel: "%EMBEDDING_MODEL_DEPLOYMENT_NAME%"
});

app.http('PromptFile', {
    methods: ['POST'],
    authLevel: 'function',
    extraInputs: [semanticSearchInput],
    handler: async (_request, context) => {
        var responseBody = context.extraInputs.get(semanticSearchInput)

        return { status: 200, body: responseBody.Response.trim() }
    }
});

const semanticSearchInput = input.generic({
    type: "semanticSearch",
    connectionName: "AISearchEndpoint",
    collection: "openai-index",
    query: "{prompt}",
    chatModel: "%CHAT_MODEL_DEPLOYMENT_NAME%",
    embeddingsModel: "%EMBEDDING_MODEL_DEPLOYMENT_NAME%"
});

app.http('PromptFile', {
    methods: ['POST'],
    authLevel: 'function',
    extraInputs: [semanticSearchInput],
    handler: async (_request, context) => {
        var responseBody: any = context.extraInputs.get(semanticSearchInput)

        return { status: 200, body: responseBody.Response.trim() }
    }
});

此示例演示如何对文件执行语义搜索。

下面是提示文件的 function.json 文件：

{
  "bindings": [
    {
      "authLevel": "function",
      "type": "httpTrigger",
      "direction": "in",
      "name": "Request",
      "methods": [
        "post"
      ]
    },
    {
      "type": "http",
      "direction": "out",
      "name": "Response"
    },
    {
      "name": "SemanticSearchInput",
      "type": "semanticSearch",
      "direction": "in",
      "connectionName": "AISearchEndpoint",
      "collection": "openai-index",
      "query": "{prompt}",
      "chatModel": "%CHAT_MODEL_DEPLOYMENT_NAME%",
      "embeddingsModel": "%EMBEDDING_MODEL_DEPLOYMENT_NAME%"
    }
  ]
}

有关 function.json 文件属性的详细信息，请参阅配置部分。

using namespace System.Net

param($Request, $TriggerMetadata, $SemanticSearchInput)

Push-OutputBinding -Name Response -Value ([HttpResponseContext]@{
        StatusCode = [HttpStatusCode]::OK
        Body       = $SemanticSearchInput.Response
    })

此示例演示如何对文件执行语义搜索。

@app.function_name("PromptFile")
@app.route(methods=["POST"])
@app.semantic_search_input(
    arg_name="result",
    search_connection_name="AISearchEndpoint",
    collection="openai-index",
    query="{prompt}",
    embeddings_model="%EMBEDDING_MODEL_DEPLOYMENT_NAME%",
    chat_model="%CHAT_MODEL_DEPLOYMENT_NAME%",
)
def prompt_file(req: func.HttpRequest, result: str) -> func.HttpResponse:
    result_json = json.loads(result)
    response_json = {
        "content": result_json.get("Response"),
        "content_type": "text/plain",
    }
    return func.HttpResponse(
        json.dumps(response_json), status_code=200, mimetype="application/json"
    )

特性

应用特性 SemanticSearchInput 来定义语义搜索输入绑定，该绑定支持以下参数：

批注

通过 SemanticSearchInput 批注可以定义语义搜索输入绑定，该绑定支持以下参数：

修饰器

在预览期间，将输入绑定定义为 generic_input_binding 类型的 semanticSearch 绑定，该绑定支持以下参数：

配置

绑定支持在 function.json 文件中设置的这些配置属性。

配置

绑定支持以下属性，这些属性在代码中定义：