Files
agent-framework/dotnet/tests/Microsoft.Agents.AI.Foundry.Hosting.UnitTests/HostedOutboundUserAgentTests.cs
Roger Barreto b12109b7e4 .NET: Bump MEAI to 10.5.1 and add Foundry per-call x-client header support (#5652)
* Bump MEAI to 10.5.1 and add per-call x-client header support

Replaces the brittle UserAgentResponsesClient subclass with a clean
per-call x-client-* header pipeline built on the new Microsoft.Extensions.AI
10.5.1 OpenAIRequestPolicies hook.

Public surface (Microsoft.Agents.AI.Foundry, [Experimental(MAAI001)]):
* chatOptions.WithClientHeader(name, value) and .WithClientHeaders(IEnumerable)
  validate the x-client- prefix (case-insensitive), apply all-or-nothing on
  bulk, and throw InvalidOperationException on foreign-typed slot collision
* myAgent.AsBuilder().UseClientHeaders().Build() opts a customer-built agent
  into the pipeline; idempotent via agent.GetService<ClientHeadersAgent>()
* Foundry-built agents (FoundryAgent.Create*) pre-wire automatically

Internals:
* ClientHeadersAgent decorator snapshots the dict at scope-push time so
  concurrent runs sharing a ChatOptions reference do not leak headers
* ClientHeadersScope is an AsyncLocal<IReadOnlyDictionary<string,string>?>
  with LIFO push/dispose semantics
* ClientHeadersPolicy singleton stamps headers via Headers.Set so per-call
  values overwrite any same-name header from earlier policies and so
  duplicate registration is value-stable
* OpenAIRequestPoliciesReflection dedups against MEAI's private _entries
  field and falls back to AddPolicy on any reflection failure; a CI test
  asserts the field shape on every MEAI bump

Hosting cleanup:
* Deleted UserAgentResponsesClient and its dummy throwing pipeline
* HostedAgentUserAgentPolicy is now registered via OpenAIRequestPolicies
  in FoundryHostingExtensions.TryApplyUserAgent

Tests:
* 19 new unit tests in ClientHeadersExtensionsTests.cs covering validation,
  AsyncLocal isolation, snapshot semantics, end-to-end wire stamping, and
  shared-chat-client dedup
* Updated OpenTelemetryAgentTests for MEAI 10.5.1 changes to web_search
  serialization and the reduced tool definition payload when sensitive
  data capture is disabled

Microsoft.Extensions.Compliance.Abstractions stays at 10.5.0 because no
10.5.1 release exists on nuget.org.

* Address PR review: pre-wire AsAIAgent path and dedup TryApplyUserAgent

* FoundryAgent: extract WireClientHeaders helper and call it from the
  internal (AIProjectClient, ChatClientAgent) constructor used by
  AzureAIProjectChatClientExtensions.AsAIAgent so those Foundry-built
  agents also pre-wire the x-client header pipeline.
* Foundry.Hosting TryApplyUserAgent: dedup HostedAgentUserAgentPolicy
  registration per OpenAIRequestPolicies instance via
  ConditionalWeakTable so per-request resolution does not grow the
  policy list unboundedly on singleton agents.

* Add tests covering AsAIAgent pre-wire and TryApplyUserAgent dedup

Backs the PR review fixes from a4c8f91 with regression tests:
* ClientHeadersExtensionsTests: AsAIAgent_FoundryAgent_HasPreWiredClientHeadersAgent
  asserts the FoundryAgent built via AzureAIProjectChatClientExtensions.AsAIAgent
  contains a ClientHeadersAgent in its delegating chain (catches future
  regressions of the bypass).
* ClientHeadersExtensionsTests: FoundryAgent_PublicConstructor_HasPreWiredClientHeadersAgent
  covers the public constructor path the same way.
* ClientHeadersExtensionsTests: UseClientHeaders_RepeatedRegistrations_OnSameChatClient_OnlyRegistersOnce
  invokes UseClientHeaders 25 times on a shared chat client and asserts via
  reflection that OpenAIRequestPolicies._entries length is exactly 1.
* HostedTryApplyUserAgentDedupTests: two tests asserting
  FoundryHostingExtensions.TryApplyUserAgent stays at one entry per
  OpenAIRequestPolicies instance after 50 calls on the same agent and across
  distinct agents on different chat clients.

* Move tests next to their SUT

Removes the dedicated HostedTryApplyUserAgentDedupTests.cs test class.
Tests are co-located with the SUT they exercise:

* FoundryAgentTests.cs gains the Constructor_PreWiresClientHeadersAgent
  and Constructor_FromAsAIAgentExtension_PreWiresClientHeadersAgent
  cases, since FoundryAgent is the SUT for the pre-wire behavior.
* HostedOutboundUserAgentTests.cs gains the two TryApplyUserAgent dedup
  cases, since FoundryHostingExtensions.TryApplyUserAgent is the SUT
  it already covers.
* ClientHeadersExtensionsTests.cs keeps only the
  UseClientHeaders_RepeatedRegistrations_OnSameChatClient_OnlyRegistersOnce
  case, which exercises the public ClientHeadersExtensions surface.

* Remove redundant WithCancellation on inner streaming call

ct is already passed to InnerAgent.RunStreamingAsync, so
.WithCancellation(ct) on the resulting IAsyncEnumerable is a no-op.
Caught by Sergey on PR review.

* Address PR review: surface downstream MEAI experimental ID

* Add AIOpenAIRequestPolicies = MEAIExperiments alias to
  DiagnosticIds.Experiments (matches the existing AIResponseContinuations,
  AIMcpServers, AIFunctionApprovals pattern).
* Mark public ClientHeadersExtensions with [Experimental(AIOpenAIRequestPolicies)]
  instead of AgentsAIExperiments. Consumers now see the MEAI001 warning,
  surfacing the dependency on MEAI's experimental OpenAIRequestPolicies hook.
* Mark internal OpenAIRequestPoliciesReflection with the same alias to
  suppress warnings at the source rather than via project-wide NoWarn.
* Remove MEAI001 from Foundry csproj NoWarn (kept on Foundry.Hosting where
  pre-PR usages remain).
* Clarify ClientHeadersScope XML doc: AsyncLocal flows values forward but
  does NOT auto-restore on method return; explicit using/Dispose is what
  gives stack-style LIFO semantics.
2026-05-06 14:43:08 +00:00

233 lines
9.6 KiB
C#

// Copyright (c) Microsoft. All rights reserved.
using System;
using System.ClientModel;
using System.ClientModel.Primitives;
using System.Collections.Generic;
using System.Net;
using System.Net.Http;
using System.Text;
using System.Threading;
using System.Threading.Tasks;
using Azure.AI.Extensions.OpenAI;
using Microsoft.AspNetCore.Builder;
using Microsoft.AspNetCore.Hosting.Server;
using Microsoft.AspNetCore.TestHost;
using Microsoft.Extensions.AI;
using Microsoft.Extensions.DependencyInjection;
using OpenAI;
#pragma warning disable OPENAI001, SCME0001, SCME0002, MEAI001
namespace Microsoft.Agents.AI.Foundry.Hosting.UnitTests;
/// <summary>
/// End-to-end tests that exercise the FULL hosted ASP.NET Core pipeline:
/// inbound HTTP → MapFoundryResponses → AgentFrameworkResponseHandler → TryApplyUserAgent →
/// agent invocation → outbound HTTP from inside the hosted environment.
/// Verifies that the hosted-agent <c>User-Agent</c> supplement reaches the outbound wire,
/// not just the inbound request.
/// </summary>
public sealed class HostedOutboundUserAgentTests : IAsyncDisposable
{
private const string TestEndpoint = "https://fake-foundry.example.com/api/projects/fake-prj";
private const string Deployment = "fake-deployment";
private WebApplication? _app;
private HttpClient? _inboundClient;
private RecordingHandler? _outboundHandler;
public async ValueTask DisposeAsync()
{
this._inboundClient?.Dispose();
this._outboundHandler?.Dispose();
if (this._app is not null)
{
await this._app.DisposeAsync();
}
}
[Fact]
public async Task Hosted_InboundResponsesRequest_TriggersOutboundCall_WithFoundryHostingSupplementAsync()
{
// Arrange: spin up a real ASP.NET Core TestServer that hosts an AIAgent backed by MEAI's
// OpenAIResponsesChatClient → ProjectResponsesClient → fake HTTP transport. This is the
// exact production stack minus the network: the only thing not real is the wire transport.
await this.StartHostedServerAsync();
// Act: send an inbound /openai/v1/responses request as the Foundry runtime would.
using var inboundRequest = new HttpRequestMessage(HttpMethod.Post, "/responses")
{
Content = new StringContent(InboundResponsesRequestJson(), Encoding.UTF8, "application/json"),
};
using var inboundResponse = await this._inboundClient!.SendAsync(inboundRequest);
var inboundBody = await inboundResponse.Content.ReadAsStringAsync();
// Assert: at least one OUTBOUND request reached the fake transport, AND it carries the
// foundry-hosting/agent-framework-dotnet/{version} supplement on its User-Agent.
// (We don't care about the inbound response shape — only that the agent's call to MEAI
// triggered an outbound request whose UA reaches the sandbox boundary correctly.)
Assert.True(this._outboundHandler!.Requests.Count > 0,
$"Expected at least one outbound request. Inbound status: {(int)inboundResponse.StatusCode}, body: {inboundBody}");
var outbound = this._outboundHandler.Requests[0];
Assert.StartsWith(TestEndpoint, outbound.Uri);
Assert.Contains("MEAI/", outbound.UserAgent);
Assert.Contains("foundry-hosting/agent-framework-dotnet", outbound.UserAgent);
}
private async Task StartHostedServerAsync()
{
var builder = WebApplication.CreateBuilder();
builder.WebHost.UseTestServer();
// Build a real ChatClientAgent whose IChatClient is MEAI's OpenAIResponsesChatClient
// wrapping a ProjectResponsesClient backed by a fake HTTP handler. After AgentFrameworkResponseHandler
// resolves this agent, TryApplyUserAgent will swap the inner _responseClient with our wrapper.
this._outboundHandler = new RecordingHandler(MinimalResponseJson());
#pragma warning disable CA5399
var outboundHttpClient = new HttpClient(this._outboundHandler);
#pragma warning restore CA5399
var projectOptions = new ProjectResponsesClientOptions
{
Transport = new HttpClientPipelineTransport(outboundHttpClient),
};
var projectResponsesClient = new ProjectResponsesClient(
new Uri(TestEndpoint),
new FakeAuthenticationTokenProvider(),
projectOptions);
IChatClient chatClient = projectResponsesClient.AsIChatClient(Deployment);
AIAgent agent = new ChatClientAgent(chatClient);
builder.Services.AddFoundryResponses(agent);
builder.Services.AddLogging();
this._app = builder.Build();
this._app.MapFoundryResponses();
await this._app.StartAsync();
var testServer = this._app.Services.GetRequiredService<IServer>() as TestServer
?? throw new InvalidOperationException("TestServer not found");
this._inboundClient = testServer.CreateClient();
}
private static string InboundResponsesRequestJson() => """
{
"model": "fake-deployment",
"input": [
{
"type": "message",
"id": "msg_1",
"status": "completed",
"role": "user",
"content": [{ "type": "input_text", "text": "Hello" }]
}
]
}
""";
private static string MinimalResponseJson() => """
{
"id":"resp_1","object":"response","created_at":1700000000,"status":"completed",
"model":"fake","output":[],"usage":{"input_tokens":1,"output_tokens":1,"total_tokens":2}
}
""";
[Fact]
public void TryApplyUserAgent_RepeatedCalls_OnSameAgent_RegistersPolicyOnce()
{
// Arrange: hosted resolution calls TryApplyUserAgent on every request. Without per-instance
// dedup, each call would append another policy entry to the shared OpenAIRequestPolicies,
// producing unbounded growth on singleton agents (one chat client reused across requests).
using var http = new HttpClient(new NoopHandler());
var openAIClient = new OpenAIClient(new ApiKeyCredential("fake"),
new OpenAIClientOptions { Transport = new HttpClientPipelineTransport(http) });
IChatClient chatClient = openAIClient.GetResponsesClient().AsIChatClient();
AIAgent agent = new ChatClientAgent(chatClient);
// Act
for (int i = 0; i < 50; i++)
{
FoundryHostingExtensions.TryApplyUserAgent(agent);
}
// Assert: exactly one HostedAgentUserAgentPolicy entry on the shared OpenAIRequestPolicies.
var policies = chatClient.GetService<OpenAIRequestPolicies>();
Assert.NotNull(policies);
Assert.Equal(1, EntriesCount(policies!));
}
[Fact]
public void TryApplyUserAgent_AcrossDistinctAgents_RegistersPolicyOncePerChatClient()
{
// Arrange: dedup is per-OpenAIRequestPolicies-instance, not global, so two agents on
// different chat clients each get exactly one registration.
using var http1 = new HttpClient(new NoopHandler());
using var http2 = new HttpClient(new NoopHandler());
var client1 = new OpenAIClient(new ApiKeyCredential("k1"),
new OpenAIClientOptions { Transport = new HttpClientPipelineTransport(http1) });
var client2 = new OpenAIClient(new ApiKeyCredential("k2"),
new OpenAIClientOptions { Transport = new HttpClientPipelineTransport(http2) });
IChatClient cc1 = client1.GetResponsesClient().AsIChatClient();
IChatClient cc2 = client2.GetResponsesClient().AsIChatClient();
AIAgent a1 = new ChatClientAgent(cc1);
AIAgent a2 = new ChatClientAgent(cc2);
// Act
for (int i = 0; i < 10; i++)
{
FoundryHostingExtensions.TryApplyUserAgent(a1);
FoundryHostingExtensions.TryApplyUserAgent(a2);
}
// Assert
Assert.Equal(1, EntriesCount(cc1.GetService<OpenAIRequestPolicies>()!));
Assert.Equal(1, EntriesCount(cc2.GetService<OpenAIRequestPolicies>()!));
}
private static int EntriesCount(OpenAIRequestPolicies policies)
{
var field = typeof(OpenAIRequestPolicies).GetField("_entries", System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.NonPublic);
var array = (Array?)field?.GetValue(policies);
return array?.Length ?? -1;
}
private sealed class NoopHandler : HttpMessageHandler
{
protected override Task<HttpResponseMessage> SendAsync(HttpRequestMessage request, CancellationToken cancellationToken)
=> Task.FromResult(new HttpResponseMessage(HttpStatusCode.OK));
}
private sealed class RecordingHandler : HttpClientHandler
{
private readonly string _body;
public List<RecordedRequest> Requests { get; } = [];
public RecordingHandler(string body)
{
this._body = body;
}
protected override Task<HttpResponseMessage> SendAsync(HttpRequestMessage request, CancellationToken cancellationToken)
{
string ua = request.Headers.TryGetValues("User-Agent", out var values)
? string.Join(",", values)
: "(none)";
this.Requests.Add(new RecordedRequest(request.RequestUri?.ToString() ?? "?", ua));
var resp = new HttpResponseMessage(HttpStatusCode.OK)
{
Content = new StringContent(this._body, Encoding.UTF8, "application/json"),
RequestMessage = request,
};
return Task.FromResult(resp);
}
}
private readonly record struct RecordedRequest(string Uri, string UserAgent);
}