Azure OpenAI assistant post input binding for Azure Functions

This example demonstrates the creation process, where the HTTP POST function that sends user prompts to the assistant chat bot. The response to the prompt is returned in the HTTP response.

/// <summary>
/// HTTP POST function that sends user prompts to the assistant chat bot.
/// </summary>
[Function(nameof(PostUserQuery))]
public static IActionResult PostUserQuery(
    [HttpTrigger(AuthorizationLevel.Anonymous, "post", Route = "assistants/{assistantId}")] HttpRequestData req,
    string assistantId,
    [AssistantPostInput("{assistantId}", "{Query.message}", ChatModel = "%CHAT_MODEL_DEPLOYMENT_NAME%", ChatStorageConnectionSetting = DefaultChatStorageConnectionSetting, CollectionName = DefaultCollectionName)] AssistantState state)
{
    return new OkObjectResult(state.RecentMessages.Any() ? state.RecentMessages[state.RecentMessages.Count - 1].Content : "No response returned.");
}

This example demonstrates the creation process, where the HTTP POST function that sends user prompts to the assistant chat bot. The response to the prompt is returned in the HTTP response.

/*
 * HTTP POST function that sends user prompts to the assistant chat bot.
 */ 
@FunctionName("PostUserResponse")
public HttpResponseMessage postUserResponse(
    @HttpTrigger(
        name = "req",
        methods = {HttpMethod.POST}, 
        authLevel = AuthorizationLevel.ANONYMOUS,
        route = "assistants/{assistantId}") 
        HttpRequestMessage<Optional<String>> request,
    @BindingName("assistantId") String assistantId,        
    @AssistantPost(name="newMessages", id = "{assistantId}", chatModel = "%CHAT_MODEL_DEPLOYMENT_NAME%", userMessage = "{Query.message}", chatStorageConnectionSetting = DEFAULT_CHATSTORAGE, collectionName = DEFAULT_COLLECTION) AssistantState state,
    final ExecutionContext context) {
        
        List<AssistantMessage> recentMessages = state.getRecentMessages();
        String response = recentMessages.isEmpty() ? "No response returned." : recentMessages.get(recentMessages.size() - 1).getContent();
        
        return request.createResponseBuilder(HttpStatus.OK)
            .header("Content-Type", "application/json")
            .body(response)
            .build();
}

This example demonstrates the creation process, where the HTTP POST function that sends user prompts to the assistant chat bot. The response to the prompt is returned in the HTTP response.

const { app, input, output } = require("@azure/functions");

const assistantPostInput = input.generic({
    type: 'assistantPost',
    id: '{assistantId}',
    chatModel: '%CHAT_MODEL_DEPLOYMENT_NAME%',
    userMessage: '{Query.message}',
    chatStorageConnectionSetting: CHAT_STORAGE_CONNECTION_SETTING,
    collectionName: COLLECTION_NAME
})
app.http('PostUserResponse', {
    methods: ['POST'],
    route: 'assistants/{assistantId}',
    authLevel: 'anonymous',
    extraInputs: [assistantPostInput],
    handler: async (_, context) => {
        const chatState = context.extraInputs.get(assistantPostInput)
        const content = chatState.recentMessages[0].content
        return {
            status: 200,
            body: content,
            headers: {
                'Content-Type': 'text/plain'
            }
        };
    }
})

import { HttpRequest, InvocationContext, app, input, output } from "@azure/functions"

const assistantPostInput = input.generic({
    type: 'assistantPost',
    id: '{assistantId}',
    chatModel: '%CHAT_MODEL_DEPLOYMENT_NAME%',
    userMessage: '{Query.message}',
    chatStorageConnectionSetting: CHAT_STORAGE_CONNECTION_SETTING,
    collectionName: COLLECTION_NAME
})
app.http('PostUserResponse', {
    methods: ['POST'],
    route: 'assistants/{assistantId}',
    authLevel: 'anonymous',
    extraInputs: [assistantPostInput],
    handler: async (_, context) => {
        const chatState: any = context.extraInputs.get(assistantPostInput)
        const content = chatState.recentMessages[0].content
        return {
            status: 200,
            body: content,
            headers: {
                'Content-Type': 'text/plain'
            }
        };
    }
})

This example demonstrates the creation process, where the HTTP POST function that sends user prompts to the assistant chat bot. The response to the prompt is returned in the HTTP response.

Here's the function.json file for post user query:

{
  "bindings": [
    {
      "authLevel": "function",
      "type": "httpTrigger",
      "direction": "in",
      "name": "Request",
      "route": "assistants/{assistantId}",
      "methods": [
        "post"
      ]
    },
    {
      "type": "http",
      "direction": "out",
      "name": "Response"
    },
    {
      "name": "State",
      "type": "assistantPost",
      "direction": "in",
      "dataType": "string",
      "id": "{assistantId}",
      "userMessage": "{Query.message}",
      "chatModel": "%CHAT_MODEL_DEPLOYMENT_NAME%",
      "chatStorageConnectionSetting": "AzureWebJobsStorage",
      "collectionName": "ChatState"
    }
  ]
}

For more information about function.json file properties, see the Configuration section.

using namespace System.Net

param($Request, $TriggerMetadata, $State)

$recent_message_content = "No recent messages!"

if ($State.recentMessages.Count -gt 0) {
    $recent_message_content = $State.recentMessages[0].content
}

Push-OutputBinding -Name Response -Value ([HttpResponseContext]@{
    StatusCode = [HttpStatusCode]::OK
    Body       = $recent_message_content
    Headers    = @{
        "Content-Type" = "text/plain"
    }
})

This example demonstrates the creation process, where the HTTP POST function that sends user prompts to the assistant chat bot. The response to the prompt is returned in the HTTP response.

@apis.function_name("PostUserQuery")
@apis.route(route="assistants/{assistantId}", methods=["POST"])
@apis.assistant_post_input(
    arg_name="state",
    id="{assistantId}",
    user_message="{Query.message}",
    chat_model="%CHAT_MODEL_DEPLOYMENT_NAME%",
    chat_storage_connection_setting=DEFAULT_CHAT_STORAGE_SETTING,
    collection_name=DEFAULT_CHAT_COLLECTION_NAME,
)
def post_user_response(req: func.HttpRequest, state: str) -> func.HttpResponse:
    # Parse the JSON string into a dictionary
    data = json.loads(state)

    # Extract the content of the recentMessage
    recent_message_content = data["recentMessages"][0]["content"]
    return func.HttpResponse(
        recent_message_content, status_code=200, mimetype="text/plain"
    )

Attributes

Apply the PostUserQuery attribute to define an assistant post input binding, which supports these parameters:

Annotations

The PostUserQuery annotation enables you to define an assistant post input binding, which supports these parameters:

Decorators

During the preview, define the output binding as a generic_output_binding binding of type postUserQuery, which supports these parameters:

Configuration

The binding supports these configuration properties that you set in the function.json file.

Configuration

The binding supports these properties, which are defined in your code: