add AzureAIInference client , wrapper and test

[Josephrp]

Why are these changes needed?

This PR introduces a new GithubLLM class to autogen, allowing users to leverage GitHub's inference endpoint with automatic fallback to Azure. It provides a seamless way to use GitHub's LLM capabilities within the autogen ecosystem, handling rate limits and ensuring high availability through Azure fallback.

Related Issues

Closes #3300

• align githubllm with [.Net][Feature Request]: Add AutoGen.AzureAIInference #3323 (comment) AzureEndpoint naming and conventions

Checks

• I've included any doc changes needed for https://microsoft.github.io/autogen/. See https://microsoft.github.io/autogen/docs/Contribute#documentation to build and test documentation locally.
• I've added tests (if relevant) corresponding to the changes introduced in this PR.
• I've made sure all auto checks have passed.

Tasks

• EmbeddingsClient
• ImageEmbeddingsClient
• ChatCompletionsClient
• EmbeddingsClient
• ImageEmbeddingsClient
• Inference

Tests

• AssistantMessage
• AsyncStreamingChatCompletions
• ChatChoice
• ChatCompletions
• ChatCompletionsFunctionToolCall
• ChatCompletionsFunctionToolDefinition
• ChatCompletionsFunctionToolSelection
• ChatCompletionsNamedFunctionToolSelection
• ChatCompletionsNamedToolSelection
• ChatCompletionsResponseFormat
• ChatCompletionsToolCall
• ChatCompletionsToolDefinition
• ChatCompletionsToolSelectionPreset
• ChatRequestMessage
• ChatResponseMessage
• ChatRole
• CompletionsFinishReason
• CompletionsUsage
• ContentItem
• EmbeddingEncodingFormat
• EmbeddingInput
• EmbeddingInputType
• EmbeddingItem
• EmbeddingsResult
• EmbeddingsUsage
• FunctionCall
• FunctionDefinition
• ImageContentItem
• ImageDetailLevel
• ImageUrl
• ModelInfo
• ModelType
• StreamingChatChoiceUpdate
• StreamingChatCompletions
• StreamingChatCompletionsUpdate
• SystemMessage
• TextContentItem
• ToolMessage
• UserMessage

Refs

• AzureAIInference
• AzureAutogenRef

[Josephrp]

@microsoft-github-policy-service agree company="Tonic-AI"

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 0% with 106 lines in your changes missing coverage. Please review.

Project coverage is 20.44%. Comparing base (6279247) to head (1e4a8c0).
Report is 70 commits behind head on main.

Files	Patch %	Lines
autogen/oai/github.py	0.00%	104 Missing and 2 partials ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #3301       +/-   ##
===========================================
- Coverage   32.90%   20.44%   -12.47%     
===========================================
  Files          94      104       +10     
  Lines       10235    11005      +770     
  Branches     2193     2501      +308     
===========================================
- Hits         3368     2250     -1118     
- Misses       6580     8508     +1928     
+ Partials      287      247       -40     

Flag	Coverage Δ
unittests	20.39% <0.00%> (-12.51%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[LittleLittleCloud]

@Josephrp Does github LLM support function call

[Josephrp]

@Josephrp Does github LLM support function call

so it's actually an azure endpoint with several models, and some of these can return a JSON , but it's not capable of running commands because it doesnt have a code execution environment (my understanding)

[LittleLittleCloud]

so it's actually an azure endpoint with several models, and some of these can return a JSON

Got it, so can I inference the github model using existing configuration but replace the endpoint and token with gh model endpoint and token?

[Josephrp]

so it's actually an azure endpoint with several models, and some of these can return a JSON

Got it, so can I inference the github model using existing configuration but replace the endpoint and token with gh model endpoint and token?

normally yes , btw i 'm starting again on this item , now + these days if you would like to work togerther there are some models with special cases , and linking the class to the message classes ... :-) fun ! 🚀

[Josephrp]

moved to here : #3559