Structure completion request to maximize Prompt Caching

[brandonh-msft]

Describe the feature or improvement you are requesting

Today, the current flow of a request through to an OpenAI service relies on simple JSON-serialization of a model to encode the message to BinaryData and send it through the pipeline.

This does not maximize Prompt Caching capabilities, where the completion request should have tools, then history, then new content - in that order.
Additionally, the tools and history must be in the same order every time (suggest alpha order by tool name).

Sources:
https://learn.microsoft.com/en-us/azure/ai-services/openai/how-to/prompt-caching
https://openai.com/index/api-prompt-caching/
https://learn.microsoft.com/en-us/azure/ai-services/openai/how-to/prompt-caching#what-is-cached

Asks for BinaryData from the options:

openai-dotnet/src/Custom/Chat/ChatClient.cs

Line 196 in c49dd70

using BinaryContent content = options.ToBinaryContent();

Writes the JSON doc in non-optimal order:

openai-dotnet/src/Generated/Models/ChatCompletionOptions.Serialization.cs

Line 15 in c49dd70

void IJsonModel<ChatCompletionOptions>.Write(Utf8JsonWriter writer, ModelReaderWriterOptions options)

Uses the non-optimal serialization when constructing the BinaryData for the options

openai-dotnet/src/Generated/Models/ChatCompletionOptions.Serialization.cs

Lines 625 to 628 in c49dd70

	internal virtual BinaryContent ToBinaryContent()
	{
	return BinaryContent.Create(this, ModelSerializationExtensions.WireOptions);
	}

Additional context

microsoft/semantic-kernel#9444

[brandonh-msft]

Looks like much of this got handled in #326 so I am going to close this (at least for now)
If folks disagree feel free to comment w/ what you're seeing/not seeing and we can re-open :)