scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-28 06:23:07 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 94

Merge ea925b3a17 into f2118fc767

Browse source
This commit is contained in:
Koudo 2026-10-22 02:38:45 +00:00 • committed by GitHub
commit 88421ded95
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 40 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

40
tests/unit/litellm_core_utils/test_streaming_chunk_builder_utils.py
View file

@ -1525,6 +1525,46 @@ def test_count_reasoning_tokens_counts_visible_reasoning():
assert processor.count_reasoning_tokens(response) > 0
def test_stream_chunk_builder_includes_reasoning_in_estimated_completion_tokens() -> None:
reasoning_content: Final = "let me count the primes under thirty"
visible_content: Final = "10"
chunks: Final = [
ModelResponseStream(
id="chatcmpl-estimated-reasoning",
model="claude-opus-4-8",
choices=[
StreamingChoices(
finish_reason=None,
index=0,
delta=Delta(role="assistant", reasoning_content=reasoning_content),
)
],
),
ModelResponseStream(
id="chatcmpl-estimated-reasoning",
model="claude-opus-4-8",
choices=[
StreamingChoices(
finish_reason="stop",
index=0,
delta=Delta(content=visible_content),
)
],
),
]
response: Final = stream_chunk_builder(chunks=chunks, messages=[{"role": "user", "content": "count"}])
assert response is not None
expected_reasoning_tokens: Final = litellm.token_counter(text=reasoning_content, count_response_tokens=True)
expected_text_tokens: Final = litellm.token_counter(
model=chunks[0].model, text=visible_content, count_response_tokens=True
)
assert response.usage.completion_tokens_details.reasoning_tokens == expected_reasoning_tokens
assert response.usage.completion_tokens == expected_text_tokens + expected_reasoning_tokens
assert response.usage.total_tokens == response.usage.prompt_tokens + response.usage.completion_tokens
@pytest.mark.parametrize(
"estimated_reasoning_tokens, expected_reasoning_tokens, expected_text_tokens",
[(40, 40, 60), (250, 100, 0)],