chore(deps): update azure azure-sdk-for-net monorepo

[renovate]

This PR contains the following updates:

Package	Change	Age	Adoption	Passing	Confidence
Azure.AI.OpenAI (source)	1.0.0-beta.5 -> 1.0.0-beta.17
Azure.Storage.Blobs (source)	12.21.2 -> 12.22.0
Azure.Storage.Files.DataLake (source)	12.19.1 -> 12.20.0

---

Release Notes

Azure/azure-sdk-for-net (Azure.AI.OpenAI)

v1.0.0-beta.17

Compare Source

1.0.0-beta.17 (2024-05-03)

Features Added

• Image input support for gpt-4-turbo chat completions now works with image data in addition to internet URLs.
  Images may be now be used as gpt-4-turbo message content items via one of three constructors:
  • ChatMessageImageContent(Uri) -- the existing constructor, used for URL-based image references
  • ChatMessageImageContent(Stream,string) -- (new) used with a stream and known MIME type (like image/png)
  • ChatMessageImageContent(BinaryData,string) -- (new) used with a BinaryData instance and known MIME type
    Please see the readme example for more details.

Breaking Changes

• Public visibility of the ChatMessageImageUrl type is removed to promote more flexible use of data sources in
  ChatMessageImageContent. Code that previously created a ChatMessageImageUrl using a Uri should simply provide
  the Uri to the ChatMessageImageContent constructor directly.

v1.0.0-beta.16

Compare Source

1.0.0-beta.16 (2024-04-11)

Features Added

Audio

• GetAudioTranscription() now supports word-level timestamp granularities via AudioTranscriptionOptions:
  • The Verbose option for ResponseFormat must be used for any timing information to be populated
  • TimestampGranularityFlags accepts a combination of the .Word and .Segment granularity values in
    AudioTimestampGranularity, joined when needed via the single-pipe | operator
    • For example, TimestampGranularityFlags = AudioTimestampGranularity.Word | AudioTimestampGranularity.Segment
      will request that both word-level and segment-level timestamps are provided on the transcription result
  • If not otherwise specified, Verbose format will default to using segment-level timestamp information
  • Corresponding word-level information is found on the .Words collection of AudioTranscription, peer to the
    existing .Segments collection
  • Note that word-level timing information incurs a small amount of additional processingly latency; segment-level
    timestamps do not encounter this behavior
• GenerateSpeechFromText() can now use Wav and Pcm values from SpeechGenerationResponseFormat, these new
  options providing alternative uncompressed formats to Flac

Chat

• ChatCompletions and StreamingChatCompletionsUpdate now include the reported Model value from the response
• Log probability information is now included in StreamingChatCompletionsUpdate when logprobs are requested on
  GetChatCompletionsStreaming()
• [AOAI] Custom Blocklist information in content filter results is now represented in a more structured
  ContentFilterDetailedResults type
• [AOAI] A new IndirectAttack content filter entry is now present on content filter results for prompts

Breaking Changes

• [AOAI] AzureChatExtensionMessageContext's RequestContentFilterResults now uses the new
  ContentFilterDetailedResults type, changed from the previous IReadOnlyList<ContentFilterBlockListIdResult>. The
  previous list is now present on CustomBlockLists.Details, supplemented with a new CustomBlockLists.Filtered
  property.

Bugs Fixed

• [AOAI] An issue that sometimes caused StreamingChatCompletionUpdates from Azure OpenAI to inappropriately exclude
  top-level information like Id and CreatedAt has been addressed

v1.0.0-beta.15

Compare Source

1.0.0-beta.15 (2024-03-20)

This release targets the latest 2024-03-01-preview service API label and brings support for the Dimensions property when using new embedding models.

Features Added

• EmbeddingsOptions now includes the Dimensions property, new to Azure OpenAI's 2024-03-01-preview service API.

Bugs Fixed

• Several issues with the ImageGenerations response object being treated as writeable are fixed:
  • ImageGenerations no longer has an erroneous public constructor
  • ImageGenerations.Created no longer has a public setter
  • ImageGenerations.Data is now an IReadOnlyList instead of an IList
  • A corresponding replacement factory method for mocks is added to AzureOpenAIModelFactory

v1.0.0-beta.14

Compare Source

1.0.0-beta.14 (2023-08-09)

Breaking Changes

• Location ip on server spans will now be set using client.address tag key on
  activity instead of http.client_ip.
  (#​37707)
• Removing ServiceVersion.V2020_09_15_Preview. This is no longer in use and
  the exporter has already defaulted to the latest ServiceVersion.v2_1.
  (#​37996)
• Remove Nullable Annotations from the Exporter's public API.
  (#​37996)

Bugs Fixed

• Fixed an issue causing no telemetry if SDK Version string exceeds max length.
  (#​37807)

Other Changes

• Update OpenTelemetry dependencies
  (#​37837)
  • OpenTelemetry 1.5.1

v1.0.0-beta.13

Compare Source

1.0.0-beta.13 (2023-07-13)

Features Added

• Added ApplicationInsightsSampler to the exporter, enabling users to customize the sampling rate using the SamplingRatio property.
  (#​36972)

Other Changes

• Updated Exporter to read v1.21.0 of the OpenTelemetry Semantic Conventions attributes for HTTP.
  For more information see Semantic conventions for HTTP spans.
  (#​37464)
  (#​37357)
• Update OpenTelemetry dependencies
  (#​36859)
  • OpenTelemetry 1.5.0
• Remove metric namespace mapping.
  (#​36968)

v1.0.0-beta.12

Compare Source

1.0.0-beta.12 (2023-12-15)

Like beta.11, beta.12 is another release that brings further refinements and fixes. It remains based on the 2023-12-01-preview service API version for Azure OpenAI and does not add any new service capabilities.

Features Added

Updates for using streaming tool calls:

• A new .NET-specific StreamingToolCallUpdate type has been added to better represent streaming tool call updates
  when using chat tools.
  • This new type includes an explicit ToolCallIndex property, reflecting index in the REST schema, to allow
    resilient deserialization of parallel function tool calling.
• A convenience constructor has been added for ChatRequestAssistantMessage that can automatically populate from a prior
  ChatResponseMessage when using non-streaming chat completions.
• A public constructor has been added for ChatCompletionsFunctionToolCall to allow more intuitive reconstruction of
  ChatCompletionsToolCall instances for use in ChatRequestAssistantMessage instances made from streaming responses.

Other additions:

• To facilitate reuse of user message contents, ChatRequestUserMessage now provides a public Content property (string) as well as a public MultimodalContentItems property (IList<ChatMessageContentItem).
  • Content is the conventional plain-text content and will be populated as non-null when the a ChatRequestUserMessage() constructor accepting a string is used to instantiate the message.
  • MultimodalContentItems is the new compound content type, currently only usable with gpt-4-vision-preview, that allows hybrid use of text and image references. It will be populated when an appropriate ChatRequestUserMessage() constructor accepting a collection of ChatMessageContentItem instances is used.
  • Role is also restored to common visibility to ChatRequestUserMessage.

Breaking Changes

• The type of ToolCallUpdate on StreamingChatCompletionsUpdate has been changed from the non-streaming
  ChatCompletionsToolCall to the new StreamingToolCallUpdate type. The conversion is straightforward:
  • ToolCallUpdate.Id remains unchanged.
  • Instead of casting ToolCallUpdate to ChatCompletionsFunctionToolCall, cast it to StreamingToolCallUpdate.
  • Update cast instance use of functionToolCallUpdate.Arguments to accumulate functionToolCallUpdate.ArgumentsUpdate.
• Removed the parameterized constructor of the ChatCompletionsOptions class that only received the messages as a parameter in favor of the parameterized constructor that receives the deployment name as well. This makes it consistent with the implementation of other Options classes.
• Removed the setter of the Input property of the EmbeddingsOptions class as per the guidelines for collection properties.

Bugs fixed

• [QUERY] Azure.AI.OpenAI_1.0.0-beta.10 no longer exposes message content on base ChatRequestMessage
• [BUG] Null Reference Exception in OpenAIClient.GetChatCompletionsAsync

v1.0.0-beta.11

Compare Source

1.0.0-beta.11 (2023-12-07)

This is a fast-following bug fix update to address some of the biggest issues reported by the community. Thank you
sharing your experiences!

Breaking Changes

• The type of ChatCompletionsOptions.ToolChoice has been updated from BinaryData to a new ChatCompletionsToolChoice type. Please use ChatCompletionsToolChoice.None, ChatCompletionsToolChoice.Auto, or provide a reference to a function or function tool definition to migrate.

Bugs Fixed

• ChatCompletionsOptions.ResponseFormat now serializes correctly and will not result in "not of type 'object" errors
• ChatCompletionsOptions.FunctionCall is fixed to again work with FunctionDefinition.None and FunctionDefinition.Auto instead of resulting in not finding a named "none" or "auto" function
• ChatCompletionsOptions.ToolChoice previously defaulted to a BinaryData type and has now been corrected to use a custom ChatCompletionsToolChoice type that parallels FunctionDefinition for older function calling.

v1.0.0-beta.10

Compare Source

1.0.0-beta.10 (2023-12-06)

Following OpenAI's November Dev Day and Microsoft's 2023 Ignite conference, this update brings a slew of new
features and changes to the SDK.

Features Added

• -1106 model feature support for gpt-35-turbo and gpt-4-turbo, including use of seed, system_fingerprint,
  parallel function calling via tools, "JSON mode" for guaranteed function outputs, and more
• dall-e-3 image generation capabilities via GetImageGenerations, featuring higher model quality, automatic prompt
  revisions by gpt-4, and customizable quality/style settings
• Greatly expanded "On Your Data" capabilities in Azure OpenAI, including many new data source options and authentication
  mechanisms
• Early support for gpt-4-vision-preview, which allows the hybrid use of text and images as input to enable scenarios
  like "describe this image for me"
• Support for Azure enhancements to gpt-4-vision-preview results that include grounding and OCR features

Breaking Changes

ChatMessage changes:

• The singular ChatMessage type has been replaced by ChatRequestMessage and ChatResponseMessage, the former of
  which is an abstract, polymorphic type with concrete derivations like ChatRequestSystemMessage and
  ChatRequestUserMessage. This requires conversion from old ChatMessages into the new types. While this is
  usually a straightforward string replacement, converting a response message into a request message (e.g. when
  propagating an assistant response to continue the conversation) will require creating a new instance of the
  appropriate request message with the response message's data. See the examples for details.

Dall-e-3:

• Azure OpenAI now uses dall-e-3 model deployments for its image generation API and such a valid deployment must
  be provided into the options for the GetImageGenerations method to receive results.

Other changes

• Audio transcription and translation (via GetAudioTranscription() and GetAudioTranslation() now allow specification of an optional Filename in addition to the binary audio data. This is used purely as an identifier and does not functionally alter the transcription/translation behavior in any way.

v1.0.0-beta.9

Compare Source

1.0.0-beta.9 (2023-11-06)

Breaking Changes

This update includes a number of version-to-version breaking changes to the API.

Streaming for completions and chat completions

Streaming Completions and Streaming Chat Completions have been significantly updated to use simpler, shallower usage
patterns and data representations. The goal of these changes is to make streaming much easier to consume in common
cases while still retaining full functionality in more complex ones (e.g. with multiple choices requested).

• A new StreamingResponse<T> type is introduced that implicitly exposes an IAsyncEnumerable<T> derived from
  the underlying response.
• OpenAI.GetCompletionsStreaming() now returns a StreamingResponse<Completions> that may be directly
  enumerated over. StreamingCompletions, StreamingChoice, and the corresponding methods are removed.
• Because Chat Completions use a distinct structure for their streaming response messages, a new
  StreamingChatCompletionsUpdate type is introduced that encapsulates this update data.
• Correspondingly, OpenAI.GetChatCompletionsStreaming() now returns a
  StreamingResponse<StreamingChatCompletionsUpdate> that may be enumerated over directly.
  StreamingChatCompletions, StreamingChatChoice, and related methods are removed.
• For more information, please see
  the related pull request description as well as the
  updated snippets in the project README.

deploymentOrModelName moved to *Options.DeploymentName

deploymentOrModelName and related method parameters on OpenAIClient have been moved to DeploymentName
properties in the corresponding method options. This is intended to promote consistency across scenario,
language, and Azure/non-Azure OpenAI use.

As an example, the following:

ChatCompletionsOptions chatCompletionsOptions = new()
{
    Messages = { new(ChatRole.User, "Hello, assistant!") },
};
Response<ChatCompletions> response = client.GetChatCompletions("gpt-4", chatCompletionsOptions);

...is now re-written as:

ChatCompletionsOptions chatCompletionsOptions = new()
{
    DeploymentName = "gpt-4",
    Messages = { new(ChatRole.User, "Hello, assistant!") },
};
Response<ChatCompletions> response = client.GetChatCompletions(chatCompletionsOptions);

Consistency in complex method options type constructors

With the migration of DeploymentName into method complex options types, these options types have now been snapped to
follow a common pattern: each complex options type will feature a default constructor that allows init-style setting
of properties as well as a single additional constructor that accepts all required parameters for the corresponding
method. Existing constructors that no longer meet that "all" requirement, including those impacted by the addition of
DeploymentName, have been removed. The "convenience" constructors that represented required parameter data
differently -- for example, EmbeddingsOptions(string), have also been removed in favor of the consistent "set of
directly provide" choice.

More exhaustively, removed are:

• AudioTranscriptionOptions(BinaryData)
• AudioTranslationOptions(BinaryData)
• ChatCompletionsOptions(IEnumerable<ChatMessage>)
• CompletionsOptions(IEnumerable<string>)
• EmbeddingsOptions(string)
• EmbeddingsOptions(IEnumerable<string>)

And added as replacements are:

• AudioTranscriptionOptions(string, BinaryData)
• AudioTranslationOptions(string, BinaryData)
• ChatCompletionsOptions(string, IEnumerable<ChatMessage>)
• CompletionsOptions(string, IEnumerable<string>)
• EmbeddingsOptions(string, IEnumerable<string>)

Embeddings now represented as ReadOnlyMemory<float>

Changed the representation of embeddings (specifically, the type of the Embedding property of the EmbeddingItem class)
from IReadOnlyList<float> to ReadOnlyMemory<float> as part of a broader effort to establish consistency across the
.NET ecosystem.

SearchKey and EmbeddingKey properties replaced by SetSearchKey and SetEmbeddingKey methods

Replaced the SearchKey and EmbeddingKey properties of the AzureCognitiveSearchChatExtensionConfiguration class with
new SetSearchKey and SetEmbeddingKey methods respectively. These methods simplify the configuration of the Azure Cognitive
Search chat extension by receiving a plain string instead of an AzureKeyCredential, promote more sensible key and secret
management, and align with the Azure SDK guidelines.

v1.0.0-beta.8

Compare Source

1.0.0-beta.8 (2023-09-21)

Features Added

• Audio Transcription and Audio Translation using OpenAI Whisper models is now supported. See OpenAI's API
  reference or the Azure OpenAI
  quickstart for detailed overview and
  background information.
  • The new methods GetAudioTranscription and GetAudioTranscription expose these capabilities on OpenAIClient
  • Transcription produces text in the primary, supported, spoken input language of the audio data provided, together
    with any optional associated metadata
  • Translation produces text, translated to English and reflective of the audio data provided, together with any
    optional associated metadata
  • These methods work for both Azure OpenAI and non-Azure api.openai.com client configurations

Breaking Changes

• The underlying representation of PromptFilterResults (for Completions and ChatCompletions) has had its response
  body key changed from prompt_annotations to prompt_filter_results
• Prior versions of the Azure.AI.OpenAI library may no longer populate PromptFilterResults as expected and it's
  highly recommended to upgrade to this version if the use of Azure OpenAI content moderation annotations for input data
  is desired
• If a library version upgrade is not immediately possible, it's advised to use Response<T>.GetRawResponse() and manually
  extract the prompt_filter_results object from the top level of the Completions or ChatCompletions response Content
  payload

Bugs Fixed

• Support for the described breaking change for PromptFilterResults was added and this library version will now again
  deserialize PromptFilterResults appropriately
• PromptFilterResults and ContentFilterResults are now exposed on the result classes for streaming Completions and
  Chat Completions. Streaming(Chat)Completions.PromptFilterResults will report an index-sorted list of all prompt
  annotations received so far while Streaming(Chat)Choice.ContentFilterResults will reflect the latest-received
  content annotations that were populated and received while streaming

v1.0.0-beta.7

Compare Source

1.0.0-beta.7 (2023-08-25)

Features Added

• The Azure OpenAI "using your own data" feature is now supported. See the Azure OpenAI using your own data quickstart for conceptual background and detailed setup instructions.
  • Azure OpenAI chat extensions are configured via a new AzureChatExtensionsOptions property on ChatCompletionsOptions. When an AzureChatExtensionsOptions is provided, configured requests will only work with clients configured to use the Azure OpenAI service, as the capabilities are unique to that service target.
  • AzureChatExtensionsOptions then has AzureChatExtensionConfiguration instances added to its Extensions property, with these instances representing the supplementary information needed for Azure OpenAI to use desired data sources to supplement chat completions behavior.
  • ChatChoice instances on a ChatCompletions response value that used chat extensions will then also have their Message property supplemented by an AzureChatExtensionMessageContext instance. This context contains a collection of supplementary Messages that describe the behavior of extensions that were used and supplementary response data, such as citations, provided along with the response.
  • See the README sample snippet for a simplified example of request/response use with "using your own data"

v1.0.0-beta.6

Compare Source

1.0.0-beta.6 (2023-07-19)

Features Added

• DALL-E image generation is now supported. See the Azure OpenAI quickstart for conceptual background and detailed setup instructions.
  • OpenAIClient gains a new GetImageGenerations method that accepts an ImageGenerationOptions and produces an ImageGenerations via its response. This response object encapsulates the temporary storage location of generated images for future retrieval.
  • In contrast to other capabilities, DALL-E image generation does not require explicit creation or specification of a deployment or model. Its surface as such does not include this concept.
• Functions for chat completions are now supported: see OpenAI's blog post on the topic for much more detail.
  • A list of FunctionDefinition objects may be populated on ChatCompletionsOptions via its Functions property. These definitions include a name and description together with a serialized JSON Schema representation of its parameters; these parameters can be generated easily via BinaryData.FromObjectAsJson with dynamic objects -- see the README for example usage.
  • NOTE: Chat Functions requires a minimum of the -0613 model versions for gpt-4 and gpt-3.5-turbo/gpt-35-turbo. Please ensure you're using these later model versions, as Functions are not supported with older model revisions. For Azure OpenAI, you can update a deployment's model version or create a new model deployment with an updated version via the Azure AI Studio interface, also accessible through Azure Portal.
• (Azure OpenAI specific) Completions and Chat Completions responses now include embedded content filter annotations for prompts and responses
• A new Azure.AI.OpenAI.AzureOpenAIModelFactory is now present for mocking.

Breaking Changes

• ChatMessage's one-parameter constructor has been replaced with a no-parameter constructor. Please replace any hybrid construction with one of these two options that either completely rely on property setting or completely rely on constructor parameters.

---

Configuration

📅 Schedule: Branch creation - At any time (no schedule defined), Automerge - At any time (no schedule defined).

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.

👻 Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.

[GoetzGoerisch]

breaking change in package
https://github.com/Azure/azure-sdk-for-net/blob/2bdb95138597211c73cb97a9cdc43606dffefec1/sdk/openai/Azure.AI.OpenAI/CHANGELOG.md#breaking-changes-5