upgrade azure ai proj

src/api/main.py

@@ -1,130 +1,147 @@
-# Copyright (c) Microsoft. All rights reserved.
-# Licensed under the MIT license.
-# See LICENSE file in the project root for full license information.
-import contextlib
-import logging
-import os
-from typing import Union
-
-import fastapi
-from azure.ai.projects.aio import AIProjectClient
-from azure.identity import AzureDeveloperCliCredential, ManagedIdentityCredential
-from dotenv import load_dotenv
-from fastapi.staticfiles import StaticFiles
-
-from .search_index_manager import SearchIndexManager
-from .util import get_logger
-
-logger = None
-enable_trace = False
-
-@contextlib.asynccontextmanager
-async def lifespan(app: fastapi.FastAPI):
-    azure_credential: Union[AzureDeveloperCliCredential, ManagedIdentityCredential]
-    if not os.getenv("RUNNING_IN_PRODUCTION"):
-        if tenant_id := os.getenv("AZURE_TENANT_ID"):
-            logger.info("Using AzureDeveloperCliCredential with tenant_id %s", tenant_id)
-            azure_credential = AzureDeveloperCliCredential(tenant_id=tenant_id)
-        else:
-            logger.info("Using AzureDeveloperCliCredential")
-            azure_credential = AzureDeveloperCliCredential()
-    else:
-        # User-assigned identity was created and set in api.bicep
-        user_identity_client_id = os.getenv("AZURE_CLIENT_ID")
-        logger.info("Using ManagedIdentityCredential with client_id %s", user_identity_client_id)
-        azure_credential = ManagedIdentityCredential(client_id=user_identity_client_id)
-
-    project = AIProjectClient(
-        credential=azure_credential,
-        endpoint=os.environ["AZURE_EXISTING_AIPROJECT_ENDPOINT"],
-    )
-
-    if enable_trace:
-        application_insights_connection_string = ""
-        try:
-            application_insights_connection_string = await project.telemetry.get_connection_string()
-        except Exception as e:
-            e_string = str(e)
-            logger.error("Failed to get Application Insights connection string, error: %s", e_string)
-        if not application_insights_connection_string:
-            logger.error("Application Insights was not enabled for this project.")
-            logger.error("Enable it via the 'Tracing' tab in your AI Foundry project page.")
-            exit()
-        else:
-            from azure.monitor.opentelemetry import configure_azure_monitor
-            configure_azure_monitor(connection_string=application_insights_connection_string)
-
-    chat = project.inference.get_chat_completions_client()
-    embed = project.inference.get_embeddings_client()
-
-    endpoint = os.environ.get('AZURE_AI_SEARCH_ENDPOINT')
-    search_index_manager = None
-    embed_dimensions = None
-    if os.getenv('AZURE_AI_EMBED_DIMENSIONS'):
-        embed_dimensions = int(os.getenv('AZURE_AI_EMBED_DIMENSIONS'))
-
-    if endpoint and os.getenv('AZURE_AI_SEARCH_INDEX_NAME') and os.getenv('AZURE_AI_EMBED_DEPLOYMENT_NAME'):
-        search_index_manager = SearchIndexManager(
-            endpoint = endpoint,
-            credential = azure_credential,
-            index_name = os.getenv('AZURE_AI_SEARCH_INDEX_NAME'),
-            dimensions = embed_dimensions,
-            model = os.getenv('AZURE_AI_EMBED_DEPLOYMENT_NAME'),
-            embeddings_client=embed
-        )
-        # Create index and upload the documents only if index does not exist.
-        logger.info(f"Creating index {os.getenv('AZURE_AI_SEARCH_INDEX_NAME')}.")
-        await search_index_manager.ensure_index_created(
-            vector_index_dimensions=embed_dimensions if embed_dimensions else 100)
-    else:
-        logger.info("The RAG search will not be used.")
-
-    app.state.chat = chat
-    app.state.search_index_manager = search_index_manager
-    app.state.chat_model = os.environ["AZURE_AI_CHAT_DEPLOYMENT_NAME"]
-    yield
-
-    await project.close()
-    await chat.close()
-    if search_index_manager is not None:
-        await search_index_manager.close()
-
-
-def create_app():
-    if not os.getenv("RUNNING_IN_PRODUCTION"):
-        load_dotenv(override=True)
-
-    global logger
-    logger = get_logger(
-        name="azureaiapp",
-        log_level=logging.INFO,
-        log_file_name = os.getenv("APP_LOG_FILE"),
-        log_to_console=True
-    )
-
-    enable_trace_string = os.getenv("ENABLE_AZURE_MONITOR_TRACING", "")
-    global enable_trace
-    enable_trace = False
-    if enable_trace_string == "":
-        enable_trace = False
-    else:
-        enable_trace = str(enable_trace_string).lower() == "true"
-    if enable_trace:
-        logger.info("Tracing is enabled.")
-        try:
-            from azure.monitor.opentelemetry import configure_azure_monitor
-        except ModuleNotFoundError:
-            logger.error("Required libraries for tracing not installed.")
-            logger.error("Please make sure azure-monitor-opentelemetry is installed.")
-            exit()
-    else:
-        logger.info("Tracing is not enabled")
-
-    app = fastapi.FastAPI(lifespan=lifespan)
-    app.mount("/static", StaticFiles(directory="api/static"), name="static")
-
-    from . import routes  # noqa
-
-    app.include_router(routes.router)
-
-    return app
+# Copyright (c) Microsoft. All rights reserved.
+# Licensed under the MIT license.
+# See LICENSE file in the project root for full license information.
+import contextlib
+import logging
+import os
+from typing import Union
+from urllib.parse import urlparse
+
+import fastapi
+from azure.ai.projects.aio import AIProjectClient
+from azure.ai.inference.aio import ChatCompletionsClient, EmbeddingsClient
+from azure.identity import AzureDeveloperCliCredential, ManagedIdentityCredential
+from dotenv import load_dotenv
+from fastapi.staticfiles import StaticFiles
+
+from .search_index_manager import SearchIndexManager
+from .util import get_logger
+
+logger = None
+enable_trace = False
+
+@contextlib.asynccontextmanager
+async def lifespan(app: fastapi.FastAPI):
+    azure_credential: Union[AzureDeveloperCliCredential, ManagedIdentityCredential]
+    if not os.getenv("RUNNING_IN_PRODUCTION"):
+        if tenant_id := os.getenv("AZURE_TENANT_ID"):
+            logger.info("Using AzureDeveloperCliCredential with tenant_id %s", tenant_id)
+            azure_credential = AzureDeveloperCliCredential(tenant_id=tenant_id)
+        else:
+            logger.info("Using AzureDeveloperCliCredential")
+            azure_credential = AzureDeveloperCliCredential()
+    else:
+        # User-assigned identity was created and set in api.bicep
+        user_identity_client_id = os.getenv("AZURE_CLIENT_ID")
+        logger.info("Using ManagedIdentityCredential with client_id %s", user_identity_client_id)
+        azure_credential = ManagedIdentityCredential(client_id=user_identity_client_id)
+
+    endpoint = os.environ["AZURE_EXISTING_AIPROJECT_ENDPOINT"]
+    project = AIProjectClient(
+        credential=azure_credential,
+        endpoint=endpoint,
+    )
+
+    if enable_trace:
+        application_insights_connection_string = ""
+        try:
+            application_insights_connection_string = await project.telemetry.get_application_insights_connection_string()
+        except Exception as e:
+            e_string = str(e)
+            logger.error("Failed to get Application Insights connection string, error: %s", e_string)
+        if not application_insights_connection_string:
+            logger.error("Application Insights was not enabled for this project.")
+            logger.error("Enable it via the 'Tracing' tab in your AI Foundry project page.")
+            exit()
+        else:
+            from azure.monitor.opentelemetry import configure_azure_monitor
+            configure_azure_monitor(connection_string=application_insights_connection_string)
+
+
+    # Project endpoint has the form:   https://your-ai-services-account-name.services.ai.azure.com/api/projects/your-project-name
+    # Inference endpoint has the form: https://your-ai-services-account-name.services.ai.azure.com/models
+    # Strip the "/api/projects/your-project-name" part and replace with "/models":
+    inference_endpoint = f"https://{urlparse(endpoint).netloc}/models"
+
+    chat =  ChatCompletionsClient(
+        endpoint=inference_endpoint,
+        credential=azure_credential,
+        credential_scopes=["https://ai.azure.com/.default"],
+    )
+    embed =  EmbeddingsClient(
+        endpoint=inference_endpoint,
+        credential=azure_credential,
+        credential_scopes=["https://ai.azure.com/.default"],
+    )
+
+    endpoint = os.environ.get('AZURE_AI_SEARCH_ENDPOINT')
+    search_index_manager = None
+    embed_dimensions = None
+    if os.getenv('AZURE_AI_EMBED_DIMENSIONS'):
+        embed_dimensions = int(os.getenv('AZURE_AI_EMBED_DIMENSIONS'))
+
+    if endpoint and os.getenv('AZURE_AI_SEARCH_INDEX_NAME') and os.getenv('AZURE_AI_EMBED_DEPLOYMENT_NAME'):
+        search_index_manager = SearchIndexManager(
+            endpoint = endpoint,
+            credential = azure_credential,
+            index_name = os.getenv('AZURE_AI_SEARCH_INDEX_NAME'),
+            dimensions = embed_dimensions,
+            model = os.getenv('AZURE_AI_EMBED_DEPLOYMENT_NAME'),
+            embeddings_client=embed
+        )
+        # Create index and upload the documents only if index does not exist.
+        logger.info(f"Creating index {os.getenv('AZURE_AI_SEARCH_INDEX_NAME')}.")
+        await search_index_manager.ensure_index_created(
+            vector_index_dimensions=embed_dimensions if embed_dimensions else 100)
+    else:
+        logger.info("The RAG search will not be used.")
+
+    app.state.chat = chat
+    app.state.search_index_manager = search_index_manager
+    app.state.chat_model = os.environ["AZURE_AI_CHAT_DEPLOYMENT_NAME"]
+    yield
+
+    await project.close()
+    await chat.close()
+    if search_index_manager is not None:
+        await search_index_manager.close()
+
+
+def create_app():
+    if not os.getenv("RUNNING_IN_PRODUCTION"):
+        load_dotenv(override=True)
+
+    global logger
+    logger = get_logger(
+        name="azureaiapp",
+        log_level=logging.INFO,
+        log_file_name = os.getenv("APP_LOG_FILE"),
+        log_to_console=True
+    )
+
+    enable_trace_string = os.getenv("ENABLE_AZURE_MONITOR_TRACING", "")
+    global enable_trace
+    enable_trace = False
+    if enable_trace_string == "":
+        enable_trace = False
+    else:
+        enable_trace = str(enable_trace_string).lower() == "true"
+    if enable_trace:
+        logger.info("Tracing is enabled.")
+        try:
+            from azure.monitor.opentelemetry import configure_azure_monitor
+        except ModuleNotFoundError:
+            logger.error("Required libraries for tracing not installed.")
+            logger.error("Please make sure azure-monitor-opentelemetry is installed.")
+            exit()
+    else:
+        logger.info("Tracing is not enabled")
+
+    app = fastapi.FastAPI(lifespan=lifespan)
+    app.mount("/static", StaticFiles(directory="api/static"), name="static")
+
+    from . import routes  # noqa
+
+    app.include_router(routes.router)
+
+    return app

src/requirements.txt

@@ -4,7 +4,7 @@ gunicorn==23.0.0
 azure-identity==1.19.0
 aiohttp==3.11.1
 azure-ai-inference==1.0.0b9
-azure-ai-projects==1.0.0.b11
+azure-ai-projects==1.0.0
 azure-search-documents
 
 azure-core==1.34.0  # other versions might not compatible