Add AI Agents and Projects Python SDK parity

sdk/ai/azure-ai-agents/CHANGELOG.md

@@ -9,12 +9,16 @@
   responses to Azure SDK `IterableStream` and Reactor `Flux` types.
 - Added raw JSON WebSocket sends, complete unknown-event payloads, UTF-8 binary JSON reception, configurable receive
   limits and overflow policies, and opt-in recovery from malformed events.
+- Added saved-job polling resumption for memory updates and agent optimization jobs.
 - Added custom WebSocket close codes and reasons, and per-event synchronous receive timeouts.
+- Added synchronous and asynchronous OpenAI factory overloads accepting a native OpenAI options callback for URL, credential, headers, query parameters, and transport overrides.
+- Added opt-in HTTP logging defaults through `AZURE_AI_PROJECTS_CONSOLE_LOGGING` and chunk-as-consumed SSE body logging in the OpenAI bridge, using the configured Java logging backend.
 - Added realtime handshake options for session IDs, structured inputs, API versions, credential scopes, preview features, extra headers and query parameters, and same-host secure connection URL overrides.
 
 - Added preview `BetaVoiceAgentsTelephonyClient` and `BetaVoiceAgentsTelephonyAsyncClient` for outbound call jobs and campaign management, including recipient import, validation, publishing, pausing, resuming, and cancellation.
 - Added preview `BetaVoiceAgentsConversationsClient` and `BetaVoiceAgentsConversationsAsyncClient` for managing
   persisted voice-agent conversations and their responses, items, and audio content.
+- Added session-affinity routing configuration through `AzureCreateResponseOptions.setRoutingConfig(...)`, `RoutingConfiguration`, and `SessionAffinityConfiguration`, with response details exposed by `ModelRouterDetails.getSessionAffinity()`.
 - Added preview `BetaVoiceAgentWebSocketClient`, `BetaVoiceAgentWebSocketAsyncClient`,
   `BetaVoiceAgentWebSocketSessionClient`, and `BetaVoiceAgentWebSocketSessionAsyncClient` with typed realtime events,
   text and PCM16 audio input, response cancellation, function-call output, persisted-conversation options, and
@@ -32,11 +36,25 @@
 ### Bugs Fixed
 
 - Reject insecure voice-agent WebSocket URLs before token acquisition to prevent sending credentials over plaintext.
+- Native asynchronous OpenAI factories and `ResponsesAsyncClient` now retrieve Azure tokens asynchronously, including factory-supplied custom OpenAI transports.
+- Supplied empty operations with zero usage when completed memory results are omitted or null.
+- Omitted multipart request and response bodies from SDK pipeline logging.
+- Preserved UTF-8 characters split across reads when logging OpenAI SSE response bodies.
 - Made synchronous voice-agent receive-buffer overflow signaling atomic across concurrent callbacks.
+- Rejected code-upload paths without a file name with an explicit argument error.
+- Agent-scoped OpenAI clients now send agent preview features, including model router controls, when
+  `AgentsClientBuilder.allowPreview(true)` is configured, and use an overridable API-version query parameter.
+- Preserved OpenAI credential and user-agent overrides through the default Azure HTTP bridge. User-supplied pipelines retain their authentication policies.
+
+- Added Java opt-in guidance to `403 preview_feature_required` errors when preview is disabled, preserving the service response and error details.
+- Preserved explicitly supplied empty `Foundry-Features` headers instead of replacing them with automatic preview opt-ins.
 - Fixed polling for telephony operations that return the `cancelled` status spelling.
+- Fixed polling for optimization jobs that return the `cancelled` status spelling.
 
 ### Other Changes
 
+- Streamed replayable code-upload content when computing SHA-256 to avoid materializing the entire upload in memory.
+
 ## 2.5.0 (2026-09-09)
 
 ### Features Added

sdk/ai/azure-ai-agents/README.md

@@ -117,6 +117,36 @@ ResponseService responseService = responsesClient.getResponseService();
 ConversationService conversationService = openAIClient.conversations();
 ```
 
+Agent-scoped OpenAI clients opt in to agent preview features when `allowPreview(true)` is configured,
+and use the configured service version. Override the defaults with native OpenAI options:
+
+```java
+OpenAIClient agentClient = builder.buildAgentScopedOpenAIClient("agent-name", options -> options
+    .replaceHeaders("User-Agent", "my-application/1.0")
+    .replaceQueryParams("api-version", "v1"));
+```
+
+The callback is also available on the project-scoped and asynchronous OpenAI factory methods.
+It supports URL, credential, headers, query parameters, and transport options. Explicit `Foundry-Features`
+headers, including empty values and case-insensitive names, are preserved. Custom OpenAI transports bypass
+the Azure pipeline. Custom Azure pipelines retain their authentication policies, which may replace
+OpenAI credential overrides. The default bridge delegates authentication to OpenAI using the builder's
+Entra credential unless overridden.
+
+Set `AZURE_AI_PROJECTS_CONSOLE_LOGGING=true` to default the builder's HTTP logging to `BODY_AND_HEADERS`.
+Native asynchronous OpenAI clients and `ResponsesAsyncClient` retrieve Azure tokens without blocking. Supply custom
+native OpenAI transports through the factory options callback to retain this authentication. Replacing the transport
+later through native `withOptions(...)` bypasses the authentication adapter and requires an explicit native credential.
+Cancelling a native OpenAI operation's future does not guarantee cancellation of pending Azure token retrieval;
+the native client's future decorators control cancellation propagation.
+Explicit `HttpLogOptions` take precedence, including `HttpLogDetailLevel.NONE` to disable HTTP logging.
+Enable INFO output in your Java logging backend (or set `AZURE_LOG_LEVEL=information` for Azure Core's
+default logger). This option does not install console handlers or change other libraries' logging levels.
+The default OpenAI bridge logs `text/event-stream` response chunks only as the caller reads them;
+it does not pre-consume the stream. Other HTTP messages use Azure Core's logging and redaction rules.
+Custom transports and custom pipelines retain their own logging configuration. Body logs are not redacted
+and can contain prompts, responses, and other sensitive data; enable them only in a trusted environment.
+
 ### Realtime connection options
 
 Use `VoiceAgentWebSocketConnectionOptions` with the synchronous or asynchronous beta voice-agent client's
@@ -234,9 +264,9 @@ and [AgentOptimizationAsyncSample.java](https://github.com/Azure/azure-sdk-for-j
 
 ### Memory item management
 
-`BetaMemoryStoresClient` and `BetaMemoryStoresAsyncClient` manage memory stores and individual memory items. In addition to store-level operations, use `createMemory`, `updateMemory`, `listMemories`, `getMemory`, and `deleteMemory` to manage individual memories. `ListMemoriesOptions` supports filtering by scope and `MemoryItemKind`, including `MemoryItemKind.PROCEDURAL`. See `MemoryStoreItemsSample` and `MemoryStoreItemsAsyncSample` for complete examples.
+`BetaMemoryStoresClient` and `BetaMemoryStoresAsyncClient` manage memory stores and individual memory items. In addition to store-level operations, use `createMemory`, `updateMemory`, `listMemories`, `getMemory`, and `deleteMemory` to manage individual memories. `ListMemoriesOptions` supports filtering by scope and `MemoryItemKind`, including `MemoryItemKind.PROCEDURAL`. See [MemoryStoreItemsSample.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/memory/MemoryStoreItemsSample.java) and [MemoryStoreItemsAsyncSample.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/memory/MemoryStoreItemsAsyncSample.java) for complete examples.
 
-For conversational memory workflows, use `beginUpdateMemories` to extract memories from conversation items, `searchMemories` to retrieve relevant memories, and `deleteScope` to remove all memories for a scope. See `MemoryStoreAdvancedSample` and `MemoryStoreAdvancedAsyncSample` for complete synchronous and asynchronous examples.
+For conversational memory workflows, use `beginUpdateMemories` to extract memories from conversation items, `searchMemories` to retrieve relevant memories, and `deleteScope` to remove all memories for a scope. See [MemoryStoreAdvancedSample.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/memory/MemoryStoreAdvancedSample.java) and [MemoryStoreAdvancedAsyncSample.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/memory/MemoryStoreAdvancedAsyncSample.java) for complete synchronous and asynchronous examples.
 
 ### Using OpenAI's official library
 
@@ -518,7 +548,7 @@ MemorySearchPreviewTool tool = new MemorySearchPreviewTool(memoryStore.getName()
     .setUpdateDelaySeconds(1);
 ```
 
-See the full sample in [MemorySearchSync.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/tools/MemorySearchSync.java) showing how to create an agent with a memory store and use it across multiple conversations.
+See the full samples in [MemorySearchSync.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/tools/MemorySearchSync.java) and [MemorySearchAsync.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/tools/MemorySearchAsync.java), which show how to create an agent with a memory store and use it across multiple conversations.
 
 ---
 
@@ -1095,6 +1125,16 @@ All realtime examples require `FOUNDRY_PROJECT_ENDPOINT` and optionally use `FOU
 
 The live audio example requires a Java Sound-compatible microphone and speaker. It streams signed, little-endian, mono PCM16 audio at 24 kHz. These examples use WebSocket transport. Although the generated protocol models include WebRTC signaling events, the Java client does not provide a WebRTC peer connection or media implementation.
 
+### Additional end-to-end samples
+
+All agent samples use `FOUNDRY_PROJECT_ENDPOINT`. Prompt-agent samples also use `FOUNDRY_MODEL_NAME`.
+
+- **Agent lifecycle and structured inputs:** [CreateAgent.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/agents/CreateAgent.java), [GetAgent.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/agents/GetAgent.java), and [CreateResponseWithStructuredInput.java](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/CreateResponseWithStructuredInput.java).
+- **Optimization jobs:** the [optimization samples](https://github.com/Azure/azure-sdk-for-java/tree/main/sdk/ai/azure-ai-agents/src/samples/java/com/azure/ai/agents/optimization) cover SDK polling, application-managed polling, cancellation, listing, retrieval, and deletion.
+- **Advanced tools:** additional samples cover structured inputs, generated-file download, File Search streaming, non-preview Web Search, custom search, and end-to-end toolbox search.
+
+---
+
 ### Service API versions
 
 The client library targets the latest service API version by default.