fix(core): Handle empty prompt in MockLLM.stream_complete

[pco111]

Fixes #19353

Description

This pull request addresses an issue where MockLLM.stream_complete() raises a pydantic.ValidationError when called with an empty prompt ("") and max_tokens=None.

The root cause was that the internal generator for the stream would complete without yielding any CompletionResponse objects. The llm_completion_callback decorator would then attempt to build a final event with a None response, leading to the validation failure.

This fix introduces a check at the beginning of the gen_prompt inner function. If the provided prompt is empty, it now yields a single CompletionResponse with empty text and delta attributes. This ensures the generator always produces a valid stream, even for empty inputs, thus resolving the bug.

New Package?

• Yes
• No

Version Bump?

• Yes
• No

Type of Change

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

• I added a new unit test to cover this change
• I believe this change is already covered by existing unit tests

from llama_index.core.llms import MockLLM

def test_mock_llm_stream_complete_empty_prompt_no_max_tokens() -> None:
    """
    Test that MockLLM.stream_complete with an empty prompt and max_tokens=None
    does not raise a validation error.
    This test case is based on issue #19353.
    """
    llm = MockLLM(max_tokens=None)
    response_gen = llm.stream_complete("")
    
    # Consume the generator to trigger the potential error
    responses = list(response_gen)
    
    # Check that we received a single, empty response
    assert len(responses) == 1
    assert responses[0].text == ""
    assert responses[0].delta == ""

Note: A temporary, specific unit test was created to validate this fix, which passed successfully. The test was then removed to maintain a clean test suite, as its sole purpose was to confirm the resolution of this specific bug.

Suggested Checklist:

• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added Google Colab support for the newly added notebooks.
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• I ran uv run make format; uv run make lint to appease the lint gods

[pco111]

Hello @logan-markewich , thank you for your previous approval. I found that there was a workflow that could not be passed (because of the lack of direct coverage of the latest changes), so I added a test (passed). Please have a look. Thank you!