ollama[patch]: permit streaming for tool calls (#28654)

Resolves https://github.com/langchain-ai/langchain/issues/28543

Ollama recently
[released](https://github.com/ollama/ollama/releases/tag/v0.4.6) support
for streaming tool calls. Previously we would override the `stream`
parameter if tools were passed in.

Covered in standard tests here:
c1d348e95d/libs/standard-tests/langchain_tests/integration_tests/chat_models.py (L893-L897)

Before, the test generates one message chunk:
```python
[
    AIMessageChunk(
        content='',
        additional_kwargs={},
        response_metadata={
            'model': 'llama3.1',
            'created_at': '2024-12-10T17:49:04.468487Z',
            'done': True,
            'done_reason': 'stop',
            'total_duration': 525471208,
            'load_duration': 19701000,
            'prompt_eval_count': 170,
            'prompt_eval_duration': 31000000,
            'eval_count': 17,
            'eval_duration': 473000000,
            'message': Message(
                role='assistant',
                content='',
                images=None,
                tool_calls=[
                    ToolCall(
                        function=Function(name='magic_function', arguments={'input': 3})
                    )
                ]
            )
        },
        id='run-552bbe0f-8fb2-4105-ada1-fa38c1db444d',
        tool_calls=[
            {
                'name': 'magic_function',
                'args': {'input': 3},
                'id': 'b0a4dc07-7d7a-487b-bd7b-ad062c2363a2',
                'type': 'tool_call',
            },
        ],
        usage_metadata={
            'input_tokens': 170, 'output_tokens': 17, 'total_tokens': 187
        },
        tool_call_chunks=[
            {
                'name': 'magic_function',
                'args': '{"input": 3}',
                'id': 'b0a4dc07-7d7a-487b-bd7b-ad062c2363a2',
                'index': None,
                'type': 'tool_call_chunk',
            }
        ]
    )
]
```

After, it generates two (tool call in one, response metadata in
another):
```python
[
    AIMessageChunk(
        content='',
        additional_kwargs={},
        response_metadata={},
        id='run-9a3f0860-baa1-4bae-9562-13a61702de70',
        tool_calls=[
            {
                'name': 'magic_function',
                'args': {'input': 3},
                'id': '5bbaee2d-c335-4709-8d67-0783c74bd2e0',
                'type': 'tool_call',
            },
        ],
        tool_call_chunks=[
            {
                'name': 'magic_function',
                'args': '{"input": 3}',
                'id': '5bbaee2d-c335-4709-8d67-0783c74bd2e0',
                'index': None,
                'type': 'tool_call_chunk',
            },
        ],
    ),
    AIMessageChunk(
        content='',
        additional_kwargs={},
        response_metadata={
            'model': 'llama3.1',
            'created_at': '2024-12-10T17:46:43.278436Z',
            'done': True,
            'done_reason': 'stop',
            'total_duration': 514282750,
            'load_duration': 16894458,
            'prompt_eval_count': 170,
            'prompt_eval_duration': 31000000,
            'eval_count': 17,
            'eval_duration': 464000000,
            'message': Message(
                role='assistant', content='', images=None, tool_calls=None
            ),
        },
        id='run-9a3f0860-baa1-4bae-9562-13a61702de70',
        usage_metadata={
            'input_tokens': 170, 'output_tokens': 17, 'total_tokens': 187
        }
    ),
]
```
This commit is contained in:
ccurme 2024-12-10 12:54:37 -05:00 committed by GitHub
parent 704059466a
commit bc4dc7f4b1
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -306,8 +306,6 @@ class ChatOllama(BaseChatModel):
'{"location": "Pune, India", "time_of_day": "morning"}'
Tool Calling:
.. warning::
Ollama currently does not support streaming for tools
.. code-block:: python
@ -459,12 +457,9 @@ class ChatOllama(BaseChatModel):
},
)
tools = kwargs.get("tools")
default_stream = not bool(tools)
params = {
"messages": ollama_messages,
"stream": kwargs.pop("stream", default_stream),
"stream": kwargs.pop("stream", True),
"model": kwargs.pop("model", self.model),
"format": kwargs.pop("format", self.format),
"options": Options(**options_dict),
@ -472,7 +467,7 @@ class ChatOllama(BaseChatModel):
**kwargs,
}
if tools:
if tools := kwargs.get("tools"):
params["tools"] = tools
return params