github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2026-06-09 10:17:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(openai): enable stream_usage when using default base URL and client (#33205)

Browse Source
This commit is contained in:
ccurme
2025-10-06 08:56:38 -04:00
committed by GitHub
parent 90e4d944ac
commit 4e50ec4b98
12 changed files with 100 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
libs/partners/openai/tests/integration_tests/chat_models/test_azure_standard.py
View File

@@ -23,7 +23,6 @@ class TestAzureOpenAIStandard(ChatModelIntegrationTests):
"deployment_name": os.environ["AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"],
"openai_api_version": OPENAI_API_VERSION,
"azure_endpoint": OPENAI_API_BASE,
"stream_usage": True,
}
@property
@@ -83,7 +82,6 @@ class TestAzureOpenAIStandardLegacy(ChatModelIntegrationTests):
"deployment_name": os.environ["AZURE_OPENAI_LEGACY_CHAT_DEPLOYMENT_NAME"],
"openai_api_version": OPENAI_API_VERSION,
"azure_endpoint": OPENAI_API_BASE,
"stream_usage": True,
}
@property

16
libs/partners/openai/tests/integration_tests/chat_models/test_base.py
View File

@@ -200,7 +200,7 @@ def test_openai_invoke() -> None:
def test_stream() -> None:
"""Test streaming tokens from OpenAI."""
llm = ChatOpenAI()
llm = ChatOpenAI(model="gpt-4.1-mini")
full: Optional[BaseMessageChunk] = None
for chunk in llm.stream("I'm Pickle Rick"):
@@ -214,7 +214,7 @@ def test_stream() -> None:
aggregate: Optional[BaseMessageChunk] = None
chunks_with_token_counts = 0
chunks_with_response_metadata = 0
for chunk in llm.stream("Hello", stream_usage=True):
for chunk in llm.stream("Hello"):
assert isinstance(chunk.content, str)
aggregate = chunk if aggregate is None else aggregate + chunk
assert isinstance(chunk, AIMessageChunk)
@@ -281,13 +281,14 @@ async def test_astream() -> None:
assert chunks_with_token_counts == 0
assert full.usage_metadata is None
llm = ChatOpenAI(temperature=0, max_tokens=MAX_TOKEN_COUNT) # type: ignore[call-arg]
await _test_stream(llm.astream("Hello"), expect_usage=False)
llm = ChatOpenAI(model="gpt-4.1-mini", temperature=0, max_tokens=MAX_TOKEN_COUNT) # type: ignore[call-arg]
await _test_stream(llm.astream("Hello", stream_usage=False), expect_usage=False)
await _test_stream(
llm.astream("Hello", stream_options={"include_usage": True}), expect_usage=True
)
await _test_stream(llm.astream("Hello", stream_usage=True), expect_usage=True)
llm = ChatOpenAI(
model="gpt-4.1-mini",
temperature=0,
max_tokens=MAX_TOKEN_COUNT, # type: ignore[call-arg]
model_kwargs={"stream_options": {"include_usage": True}},
@@ -297,7 +298,12 @@ async def test_astream() -> None:
llm.astream("Hello", stream_options={"include_usage": False}),
expect_usage=False,
)
llm = ChatOpenAI(temperature=0, max_tokens=MAX_TOKEN_COUNT, stream_usage=True) # type: ignore[call-arg]
llm = ChatOpenAI(
model="gpt-4.1-mini",
temperature=0,
max_tokens=MAX_TOKEN_COUNT, # type: ignore[call-arg]
stream_usage=True,
)
await _test_stream(llm.astream("Hello"), expect_usage=True)
await _test_stream(llm.astream("Hello", stream_usage=False), expect_usage=False)

2
libs/partners/openai/tests/integration_tests/chat_models/test_base_standard.py
View File

@@ -22,7 +22,7 @@ class TestOpenAIStandard(ChatModelIntegrationTests):
@property
def chat_model_params(self) -> dict:
return {"model": "gpt-4o-mini", "stream_usage": True}
return {"model": "gpt-4o-mini"}
@property
def supports_image_inputs(self) -> bool: