Azure AI Inference SDK - Changes for beta 4 release (#36669)

Author: dargilco

.vscode/cspell.json

@@ -1305,7 +1305,8 @@
         "mros",
         "Nify",
         "ctxt",
-        "wday"
+        "wday",
+        "dtype"
       ]
     },
     {

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

@@ -1,14 +1,24 @@
 # Release History
 
-## 1.0.0b4 (Unreleased)
+## 1.0.0b4 (2024-08-30)
 
 ### Features Added
 
+* Support chat completion streaming response with function arguments (tool calls). Add new classes
+`StreamingChatResponseMessageUpdate` and `StreamingChatResponseToolCallUpdate`.
+* Support text embeddings result in base64 encoded string format.
+* Nicely formated print of chat completions and embeddings result objects.
+
 ### Breaking Changes
 
+* Classes `ChatCompletionsToolSelectionPreset`, `ChatCompletionsNamedToolSelection` and `ChatCompletionsFunctionToolSelection` renamed to `ChatCompletionsToolChoicePreset` `ChatCompletionsNamedToolChoice` and `ChatCompletionsNamedToolChoiceFunction` respectively.
+* Update the object type of `embeddings` property on `EmbeddingsResult`, from `embedding: List[float]` to `embedding: Union[str, List[float]]`.
+* Instead of base class `ChatCompletionsToolCall` and derived class `ChatCompletionsFunctionToolCall`, we now have a flat representation of only one class `ChatCompletionsToolCall` that that represents a function tool. This is because the only support tool is a function call.
+
 ### Bugs Fixed
 
-### Other Changes
+* Fix setting of chat completions response format, to allow response in JSON format. See classes `ChatCompletionsResponseFormat` (base class) and
+derived classes `ChatCompletionsResponseFormatJSON` and `ChatCompletionsResponseFormatText`.
 
 ## 1.0.0b3 (2024-07-31)
 

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

@@ -2,5 +2,5 @@
   "AssetsRepo": "Azure/azure-sdk-assets",
   "AssetsRepoPrefixPath": "python",
   "TagPrefix": "python/ai/azure-ai-inference",
-  "Tag": "python/ai/azure-ai-inference_a43207800c"
+  "Tag": "python/ai/azure-ai-inference_498e85cbfd"
 }

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

@@ -37,10 +37,10 @@ class ChatCompletionsClient(ChatCompletionsClientOperationsMixin):  # pylint: di
 
     :param endpoint: Service host. Required.
     :type endpoint: str
-    :param credential: Credential used to authenticate requests to the service. Is either a
-     AzureKeyCredential type or a TokenCredential type. Required.
+    :param credential: Credential used to authenticate requests to the service. Is one of the
+     following types: AzureKeyCredential, AzureKeyCredential, TokenCredential Required.
     :type credential: ~azure.core.credentials.AzureKeyCredential or
-     ~azure.core.credentials.TokenCredential
+     ~azure.core.credentials.AzureKeyCredential or ~azure.core.credentials.TokenCredential
     :keyword api_version: The API version to use for this operation. Default value is
      "2024-05-01-preview". Note that overriding this default value may result in unsupported
      behavior.
@@ -115,10 +115,10 @@ class EmbeddingsClient(EmbeddingsClientOperationsMixin):  # pylint: disable=clie
 
     :param endpoint: Service host. Required.
     :type endpoint: str
-    :param credential: Credential used to authenticate requests to the service. Is either a
-     AzureKeyCredential type or a TokenCredential type. Required.
+    :param credential: Credential used to authenticate requests to the service. Is one of the
+     following types: AzureKeyCredential, AzureKeyCredential, TokenCredential Required.
     :type credential: ~azure.core.credentials.AzureKeyCredential or
-     ~azure.core.credentials.TokenCredential
+     ~azure.core.credentials.AzureKeyCredential or ~azure.core.credentials.TokenCredential
     :keyword api_version: The API version to use for this operation. Default value is
      "2024-05-01-preview". Note that overriding this default value may result in unsupported
      behavior.
@@ -193,10 +193,10 @@ class ImageEmbeddingsClient(ImageEmbeddingsClientOperationsMixin):  # pylint: di
 
     :param endpoint: Service host. Required.
     :type endpoint: str
-    :param credential: Credential used to authenticate requests to the service. Is either a
-     AzureKeyCredential type or a TokenCredential type. Required.
+    :param credential: Credential used to authenticate requests to the service. Is one of the
+     following types: AzureKeyCredential, AzureKeyCredential, TokenCredential Required.
     :type credential: ~azure.core.credentials.AzureKeyCredential or
-     ~azure.core.credentials.TokenCredential
+     ~azure.core.credentials.AzureKeyCredential or ~azure.core.credentials.TokenCredential
     :keyword api_version: The API version to use for this operation. Default value is
      "2024-05-01-preview". Note that overriding this default value may result in unsupported
      behavior.

sdk/ai/azure-ai-inference/azure/ai/inference/_client.py

@@ -26,10 +26,10 @@ class ChatCompletionsClientConfiguration:  # pylint: disable=too-many-instance-a
 
     :param endpoint: Service host. Required.
     :type endpoint: str
-    :param credential: Credential used to authenticate requests to the service. Is either a
-     AzureKeyCredential type or a TokenCredential type. Required.
+    :param credential: Credential used to authenticate requests to the service. Is one of the
+     following types: AzureKeyCredential, AzureKeyCredential, TokenCredential Required.
     :type credential: ~azure.core.credentials.AzureKeyCredential or
-     ~azure.core.credentials.TokenCredential
+     ~azure.core.credentials.AzureKeyCredential or ~azure.core.credentials.TokenCredential
     :keyword api_version: The API version to use for this operation. Default value is
      "2024-05-01-preview". Note that overriding this default value may result in unsupported
      behavior.
@@ -55,6 +55,8 @@ def __init__(self, endpoint: str, credential: Union[AzureKeyCredential, "TokenCr
     def _infer_policy(self, **kwargs):
         if isinstance(self.credential, AzureKeyCredential):
             return policies.AzureKeyCredentialPolicy(self.credential, "Authorization", prefix="Bearer", **kwargs)
+        if isinstance(self.credential, AzureKeyCredential):
+            return policies.AzureKeyCredentialPolicy(self.credential, "api-key", **kwargs)
         if hasattr(self.credential, "get_token"):
             return policies.BearerTokenCredentialPolicy(self.credential, *self.credential_scopes, **kwargs)
         raise TypeError(f"Unsupported credential: {self.credential}")
@@ -81,10 +83,10 @@ class EmbeddingsClientConfiguration:  # pylint: disable=too-many-instance-attrib
 
     :param endpoint: Service host. Required.
     :type endpoint: str
-    :param credential: Credential used to authenticate requests to the service. Is either a
-     AzureKeyCredential type or a TokenCredential type. Required.
+    :param credential: Credential used to authenticate requests to the service. Is one of the
+     following types: AzureKeyCredential, AzureKeyCredential, TokenCredential Required.
     :type credential: ~azure.core.credentials.AzureKeyCredential or
-     ~azure.core.credentials.TokenCredential
+     ~azure.core.credentials.AzureKeyCredential or ~azure.core.credentials.TokenCredential
     :keyword api_version: The API version to use for this operation. Default value is
      "2024-05-01-preview". Note that overriding this default value may result in unsupported
      behavior.
@@ -110,6 +112,8 @@ def __init__(self, endpoint: str, credential: Union[AzureKeyCredential, "TokenCr
     def _infer_policy(self, **kwargs):
         if isinstance(self.credential, AzureKeyCredential):
             return policies.AzureKeyCredentialPolicy(self.credential, "Authorization", prefix="Bearer", **kwargs)
+        if isinstance(self.credential, AzureKeyCredential):
+            return policies.AzureKeyCredentialPolicy(self.credential, "api-key", **kwargs)
         if hasattr(self.credential, "get_token"):
             return policies.BearerTokenCredentialPolicy(self.credential, *self.credential_scopes, **kwargs)
         raise TypeError(f"Unsupported credential: {self.credential}")
@@ -136,10 +140,10 @@ class ImageEmbeddingsClientConfiguration:  # pylint: disable=too-many-instance-a
 
     :param endpoint: Service host. Required.
     :type endpoint: str
-    :param credential: Credential used to authenticate requests to the service. Is either a
-     AzureKeyCredential type or a TokenCredential type. Required.
+    :param credential: Credential used to authenticate requests to the service. Is one of the
+     following types: AzureKeyCredential, AzureKeyCredential, TokenCredential Required.
     :type credential: ~azure.core.credentials.AzureKeyCredential or
-     ~azure.core.credentials.TokenCredential
+     ~azure.core.credentials.AzureKeyCredential or ~azure.core.credentials.TokenCredential
     :keyword api_version: The API version to use for this operation. Default value is
      "2024-05-01-preview". Note that overriding this default value may result in unsupported
      behavior.
@@ -165,6 +169,8 @@ def __init__(self, endpoint: str, credential: Union[AzureKeyCredential, "TokenCr
     def _infer_policy(self, **kwargs):
         if isinstance(self.credential, AzureKeyCredential):
             return policies.AzureKeyCredentialPolicy(self.credential, "Authorization", prefix="Bearer", **kwargs)
+        if isinstance(self.credential, AzureKeyCredential):
+            return policies.AzureKeyCredentialPolicy(self.credential, "api-key", **kwargs)
         if hasattr(self.credential, "get_token"):
             return policies.BearerTokenCredentialPolicy(self.credential, *self.credential_scopes, **kwargs)
         raise TypeError(f"Unsupported credential: {self.credential}")

sdk/ai/azure-ai-inference/azure/ai/inference/_configuration.py

@@ -476,6 +476,9 @@ def _create_value(rf: typing.Optional["_RestField"], value: typing.Any) -> typin
 
 class Model(_MyMutableMapping):
     _is_model = True
+    # label whether current class's _attr_to_rest_field has been calculated
+    # could not see _attr_to_rest_field directly because subclass inherits it from parent class
+    _calculated: typing.Set[str] = set()
 
     def __init__(self, *args: typing.Any, **kwargs: typing.Any) -> None:
         class_name = self.__class__.__name__
@@ -508,24 +511,27 @@ def copy(self) -> "Model":
         return Model(self.__dict__)
 
     def __new__(cls, *args: typing.Any, **kwargs: typing.Any) -> Self:  # pylint: disable=unused-argument
-        # we know the last three classes in mro are going to be 'Model', 'dict', and 'object'
-        mros = cls.__mro__[:-3][::-1]  # ignore model, dict, and object parents, and reverse the mro order
-        attr_to_rest_field: typing.Dict[str, _RestField] = {  # map attribute name to rest_field property
-            k: v for mro_class in mros for k, v in mro_class.__dict__.items() if k[0] != "_" and hasattr(v, "_type")
-        }
-        annotations = {
-            k: v
-            for mro_class in mros
-            if hasattr(mro_class, "__annotations__")  # pylint: disable=no-member
-            for k, v in mro_class.__annotations__.items()  # pylint: disable=no-member
-        }
-        for attr, rf in attr_to_rest_field.items():
-            rf._module = cls.__module__
-            if not rf._type:
-                rf._type = rf._get_deserialize_callable_from_annotation(annotations.get(attr, None))
-            if not rf._rest_name_input:
-                rf._rest_name_input = attr
-        cls._attr_to_rest_field: typing.Dict[str, _RestField] = dict(attr_to_rest_field.items())
+        if f"{cls.__module__}.{cls.__qualname__}" not in cls._calculated:
+            # we know the last nine classes in mro are going to be 'Model', '_MyMutableMapping', 'MutableMapping',
+            # 'Mapping', 'Collection', 'Sized', 'Iterable', 'Container' and 'object'
+            mros = cls.__mro__[:-9][::-1]  # ignore parents, and reverse the mro order
+            attr_to_rest_field: typing.Dict[str, _RestField] = {  # map attribute name to rest_field property
+                k: v for mro_class in mros for k, v in mro_class.__dict__.items() if k[0] != "_" and hasattr(v, "_type")
+            }
+            annotations = {
+                k: v
+                for mro_class in mros
+                if hasattr(mro_class, "__annotations__")  # pylint: disable=no-member
+                for k, v in mro_class.__annotations__.items()  # pylint: disable=no-member
+            }
+            for attr, rf in attr_to_rest_field.items():
+                rf._module = cls.__module__
+                if not rf._type:
+                    rf._type = rf._get_deserialize_callable_from_annotation(annotations.get(attr, None))
+                if not rf._rest_name_input:
+                    rf._rest_name_input = attr
+            cls._attr_to_rest_field: typing.Dict[str, _RestField] = dict(attr_to_rest_field.items())
+            cls._calculated.add(f"{cls.__module__}.{cls.__qualname__}")
 
         return super().__new__(cls)  # pylint: disable=no-value-for-parameter
 
@@ -563,6 +569,7 @@ def as_dict(self, *, exclude_readonly: bool = False) -> typing.Dict[str, typing.
         """
 
         result = {}
+        readonly_props = []
         if exclude_readonly:
             readonly_props = [p._rest_name for p in self._attr_to_rest_field.values() if _is_readonly(p)]
         for k, v in self.items():