[JS] feat: add Streaming support

[Stevenic]

Linked issues

closes: #1593

Details

Updated JS OpenAIModel class to support streaming and updated ActionPlanner to support streaming chunks to the client.

Change details

• Added a new StreamingResponse class that implements the new activity based chunking protocol.
• Added a new optional set of events that prompt completion models can implement.
• Updated the OpenAIModel class to implement the new prompt completion model events.
• Updated the OpenAIModel class to use the official node library and added streaming support.
• Added a new StreamingLLMClient which manages the task of streaming chunks from the model to the client using the new StreamingResponse class.
• Updated the ActionPlanner class to use the new StreamingLLMClient instead of the LLMClient class directly.
• Updated the ActionPlanner to return an empty plan anytime streaming is used.
• Added a new TestModel class to simplify testing of things like streaming support.
• Added basic unit tests for the StreamingResponse, TestModel, LLMClient, and StreamingLLMClient classes. `

Attestation Checklist

• [x ] My code follows the style guidelines of this project

• I have checked for/fixed spelling, linting, and other errors

• I have commented my code for clarity

• I have made corresponding changes to the documentation (updating the doc strings in the code is sufficient)

• My changes generate no new warnings

• I have added tests that validates my changes, and provides sufficient test coverage. I have tested with:

  • Local testing
  • E2E testing in Teams

• New and existing unit tests pass locally with my changes

Additional information

I started basic unit tests for the LLMClient class. These tests should be extended to provider better coverage. We should also add unit tests for the ActionPlanner class now that the TestModel class exists.

[corinagum]

@Stevenic could you mark the class/methods in LLMClient as deprecated? Notes indicating to use StreamingLLMClient instead would be great.

[Stevenic]

@Stevenic could you mark the class/methods in LLMClient as deprecated? Notes indicating to use StreamingLLMClient instead would be great.

It's not really deprecated though. You should just generally use the derived class. It will trigger warnings if I mark it as deprecated since it's a base class

[corinagum]

It's not really deprecated though. You should just generally use the derived class. It will trigger warnings if I mark it as deprecated since it's a base class

Okay so maybe marking it as deprecated isn't the right answer, but I find it confusing that the base class is only being used by one derived class. This is specifically done for backwards compatibility, correct? In that case there should be remarks to that effect, and what scenarios you would want to use the base class. Feel free to let me know if this is just misconception from my functions oriented brain, but personally the extra comments would be beneficial.

[Stevenic]

It's not really deprecated though. You should just generally use the derived class. It will trigger warnings if I mark it as deprecated since it's a base class

Okay so maybe marking it as deprecated isn't the right answer, but I find it confusing that the base class is only being used by one derived class. This is specifically done for backwards compatibility, correct? In that case there should be remarks to that effect, and what scenarios you would want to use the base class. Feel free to let me know if this is just misconception from my functions oriented brain, but personally the extra comments would be beneficial.

I can merge the two classes