Adding test recording for pause, resume and deployment finetuning fea… (#44428)

* Adding test recording for pause, resume and deployment finetuning features

* adding islive condition

---------

Co-authored-by: Jayesh Tanna <[email protected]>

Author: jayesh-tanna

sdk/ai/azure-ai-projects/assets.json

@@ -2,5 +2,5 @@
   "AssetsRepo": "Azure/azure-sdk-assets",
   "AssetsRepoPrefixPath": "python",
   "TagPrefix": "python/ai/azure-ai-projects",
-  "Tag": "python/ai/azure-ai-projects_1d985f7ba5"
+  "Tag": "python/ai/azure-ai-projects_5b7a7fa09e"
 }

sdk/ai/azure-ai-projects/tests/conftest.py

@@ -111,6 +111,9 @@ def sanitize_url_paths():
     # Sanitize fine-tuning job IDs in URLs and response bodies
     add_general_regex_sanitizer(regex=r"ftjob-[a-f0-9]+", value="sanitized-ftjob-id")
 
+    # Sanitize deployment names that are derived from job IDs (e.g., test-6158cfe2)
+    add_general_regex_sanitizer(regex=r"test-[a-f0-9]{8}", value="test-ftjob-id")
+
     # Sanitize file IDs in URLs and response bodies
     add_general_regex_sanitizer(regex=r"file-[a-f0-9]+", value="sanitized-file-id")
 
@@ -160,3 +163,11 @@ def sanitize_url_paths():
     #  - AZSDK3430: $..id
     remove_batch_sanitizers(["AZSDK3493"])
     remove_batch_sanitizers(["AZSDK3430"])
+
+    # Sanitize ARM operation headers that contain certificates and identifiers
+    add_general_regex_sanitizer(regex=r"[?&]t=[0-9]+", value="&t=sanitized-timestamp")
+    add_general_regex_sanitizer(regex=r"[?&]c=[^&\"]+", value="&c=sanitized-certificate")
+    add_general_regex_sanitizer(regex=r"[?&]s=[^&\"]+", value="&s=sanitized-signature")
+    add_general_regex_sanitizer(regex=r"[?&]h=[^&\"]+", value="&h=sanitized-hash")
+    add_general_regex_sanitizer(regex=r"operationResults/[a-f0-9\-]+", value="operationResults/sanitized-operation-id")
+    add_general_regex_sanitizer(regex=r"https://management\.azure\.com/", value="https://sanitized.azure.com/")

sdk/ai/azure-ai-projects/tests/finetuning/test_finetuning.py

@@ -20,7 +20,7 @@
 from devtools_testutils import (
     recorded_by_proxy,
     RecordedTransport,
-    is_live_and_not_recording,
+    is_live
 )
 from azure.mgmt.cognitiveservices import CognitiveServicesManagementClient
 from azure.mgmt.cognitiveservices.models import Deployment, DeploymentProperties, DeploymentModel, Sku
@@ -346,8 +346,9 @@ def _test_deploy_and_infer_helper(
                         print(f"[{test_prefix}] Deployment status: {deployment_operation.status()}")
 
                     print(f"[{test_prefix}] Deployment completed successfully")
-                    print(f"[{test_prefix}] Waiting for 120 seconds for deployment to be fully ready.")
-                    time.sleep(120)
+                    if is_live():
+                        print(f"[{test_prefix}] Waiting for 10 minutes for deployment to be fully ready.")
+                        time.sleep(600)
                     print(f"[{test_prefix}] Testing inference on deployment: {deployment_name}")
 
                     response = openai_client.responses.create(
@@ -622,10 +623,6 @@ def test_finetuning_list_events(self, **kwargs):
                 self._cleanup_test_file(openai_client, train_file.id)
                 self._cleanup_test_file(openai_client, validation_file.id)
 
-    @pytest.mark.skipif(
-        condition=(not is_live_and_not_recording()),
-        reason="Skipped because not able to pause any job",
-    )
     @servicePreparer()
     @recorded_by_proxy(RecordedTransport.AZURE_CORE, RecordedTransport.HTTPX)
     def test_finetuning_pause_job(self, **kwargs):
@@ -658,10 +655,6 @@ def test_finetuning_pause_job(self, **kwargs):
 
                 print(f"[test_finetuning_pause_job] Successfully paused and verified job: {running_job_id}")
 
-    @pytest.mark.skipif(
-        condition=(not is_live_and_not_recording()),
-        reason="Skipped because not able to pause any job",
-    )
     @servicePreparer()
     @recorded_by_proxy(RecordedTransport.AZURE_CORE, RecordedTransport.HTTPX)
     def test_finetuning_resume_job(self, **kwargs):
@@ -733,10 +726,6 @@ def test_finetuning_list_checkpoints(self, **kwargs):
                     f"[test_finetuning_list_checkpoints] Successfully validated {len(checkpoints_list)} checkpoints for job: {completed_job_id}"
                 )
 
-    @pytest.mark.skipif(
-        condition=(not is_live_and_not_recording()),
-        reason="Skipped because API not sending completed or failed status despite job being complete",
-    )
     @servicePreparer()
     @recorded_by_proxy(RecordedTransport.AZURE_CORE, RecordedTransport.HTTPX)
     def test_finetuning_deploy_and_infer_oai_model_sft_job(self, **kwargs):
@@ -750,12 +739,8 @@ def test_finetuning_deploy_and_infer_oai_model_sft_job(self, **kwargs):
             **kwargs,
         )
 
-    @pytest.mark.skipif(
-        condition=(not is_live_and_not_recording()),
-        reason="Skipped because not able to complete any RFT job",
-    )
     @servicePreparer()
-    @recorded_by_proxy
+    @recorded_by_proxy(RecordedTransport.AZURE_CORE, RecordedTransport.HTTPX)
     def test_finetuning_deploy_and_infer_oai_model_rft_job(self, **kwargs):
         completed_job_id = kwargs.get("azure_ai_projects_tests_completed_oai_model_rft_fine_tuning_job_id")
         self._test_deploy_and_infer_helper(
@@ -767,10 +752,6 @@ def test_finetuning_deploy_and_infer_oai_model_rft_job(self, **kwargs):
             **kwargs,
         )
 
-    @pytest.mark.skipif(
-        condition=(not is_live_and_not_recording()),
-        reason="Skipped because API not sending completed or failed status despite job being complete",
-    )
     @servicePreparer()
     @recorded_by_proxy(RecordedTransport.AZURE_CORE, RecordedTransport.HTTPX)
     def test_finetuning_deploy_and_infer_oai_model_dpo_job(self, **kwargs):
@@ -784,10 +765,6 @@ def test_finetuning_deploy_and_infer_oai_model_dpo_job(self, **kwargs):
             **kwargs,
         )
 
-    @pytest.mark.skipif(
-        condition=(not is_live_and_not_recording()),
-        reason="Skipped because API not sending completed or failed status despite job being complete",
-    )
     @servicePreparer()
     @recorded_by_proxy(RecordedTransport.AZURE_CORE, RecordedTransport.HTTPX)
     def test_finetuning_deploy_and_infer_oss_model_sft_job(self, **kwargs):

sdk/ai/azure-ai-projects/tests/finetuning/test_finetuning_async.py

@@ -351,8 +351,8 @@ async def _test_deploy_and_infer_helper_async(
                     print(f"[{test_prefix}] Deployment status: {deployment_operation.status()}")
 
                 print(f"[{test_prefix}] Deployment completed successfully")
-                print(f"[{test_prefix}] Waiting for 120 seconds for deployment to be fully ready.")
-                await asyncio.sleep(120)
+                print(f"[{test_prefix}] Waiting for 10 minutes for deployment to be fully ready.")
+                await asyncio.sleep(600)
                 print(f"[{test_prefix}] Testing inference on deployment: {deployment_name}")
 
                 response = await openai_client.responses.create(
@@ -654,10 +654,6 @@ async def test_finetuning_list_events_async(self, **kwargs):
             await self._cleanup_test_file_async(openai_client, train_file.id)
             await self._cleanup_test_file_async(openai_client, validation_file.id)
 
-    @pytest.mark.skipif(
-        condition=(not is_live_and_not_recording()),
-        reason="Skipped because not able to pause any job",
-    )
     @servicePreparer()
     @recorded_by_proxy_async(RecordedTransport.AZURE_CORE, RecordedTransport.HTTPX)
     async def test_finetuning_pause_job_async(self, **kwargs):
@@ -691,10 +687,6 @@ async def test_finetuning_pause_job_async(self, **kwargs):
             print(f"[test_finetuning_pause_job] Job status after pause: {paused_job.status}")
             print(f"[test_finetuning_pause_job] Successfully paused and verified job: {running_job_id}")
 
-    @pytest.mark.skipif(
-        condition=(not is_live_and_not_recording()),
-        reason="Skipped because not able to pause any job",
-    )
     @servicePreparer()
     @recorded_by_proxy_async(RecordedTransport.AZURE_CORE, RecordedTransport.HTTPX)
     async def test_finetuning_resume_job_async(self, **kwargs):