🐫 CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents. https://www.camel-ai.org
Python agentai-societiesartificial-intelligencecommunicative-aicooperative-aideep-learninglarge-language-modelsmulti-agent-systemsnatural-language-processing
<a id="camel.models.base_model"></a>
<a id="camel.models.base_model._StreamLogger"></a>
## _StreamLogger
```python
class _StreamLogger:
```
Base for stream logging wrappers.
<a id="camel.models.base_model._StreamLogger.__init__"></a>
### __init__
```python
def __init__(self, log_path: Optional[str], log_enabled: bool):
```
<a id="camel.models.base_model._StreamLogger._collect"></a>
### _collect
```python
def _collect(self, chunk: ChatCompletionChunk):
```
<a id="camel.models.base_model._StreamLogger._log"></a>
### _log
```python
def _log(self):
```
<a id="camel.models.base_model._SyncStreamWrapper"></a>
## _SyncStreamWrapper
```python
class _SyncStreamWrapper(_StreamLogger):
```
Sync stream wrapper with logging.
<a id="camel.models.base_model._SyncStreamWrapper.__init__"></a>
### __init__
```python
def __init__(
self,
stream: Union[Stream[ChatCompletionChunk], Generator[ChatCompletionChunk, None, None]],
log_path: Optional[str],
log_enabled: bool
):
```
<a id="camel.models.base_model._SyncStreamWrapper.__iter__"></a>
### __iter__
```python
def __iter__(self):
```
<a id="camel.models.base_model._SyncStreamWrapper.__next__"></a>
### __next__
```python
def __next__(self):
```
<a id="camel.models.base_model._SyncStreamWrapper.__enter__"></a>
### __enter__
```python
def __enter__(self):
```
<a id="camel.models.base_model._SyncStreamWrapper.__exit__"></a>
### __exit__
```python
def __exit__(self, *_):
```
<a id="camel.models.base_model._SyncStreamWrapper.__del__"></a>
### __del__
```python
def __del__(self):
```
<a id="camel.models.base_model._AsyncStreamWrapper"></a>
## _AsyncStreamWrapper
```python
class _AsyncStreamWrapper(_StreamLogger):
```
Async stream wrapper with logging.
<a id="camel.models.base_model._AsyncStreamWrapper.__init__"></a>
### __init__
```python
def __init__(
self,
stream: Union[AsyncStream[ChatCompletionChunk], AsyncGenerator[ChatCompletionChunk, None]],
log_path: Optional[str],
log_enabled: bool
):
```
<a id="camel.models.base_model._AsyncStreamWrapper.__aiter__"></a>
### __aiter__
```python
def __aiter__(self):
```
<a id="camel.models.base_model._AsyncStreamWrapper.__del__"></a>
### __del__
```python
def __del__(self):
```
<a id="camel.models.base_model.ModelBackendMeta"></a>
## ModelBackendMeta
```python
class ModelBackendMeta(ABCMeta):
```
Metaclass that automatically pre/post-processes the run method.
Automatically wraps the run and arun methods of any class inheriting from
BaseModelBackend to:
- Preprocess messages (remove `<think>` tags) before sending to the model.
- Postprocess responses (extract `<think>` tags into reasoning_content)
after receiving from the model.
<a id="camel.models.base_model.ModelBackendMeta.__new__"></a>
### __new__
```python
def __new__(
mcs,
name,
bases,
namespace
):
```
Wraps run/arun methods with pre/post-processing.
<a id="camel.models.base_model.BaseModelBackend"></a>
## BaseModelBackend
```python
class BaseModelBackend(ABC):
```
Base class for different model backends.
It may be OpenAI API, a local LLM, a stub for unit tests, etc.
**Parameters:**
- **model_type** (Union[ModelType, str]): Model for which a backend is created.
- **model_config_dict** (Optional[Dict[str, Any]], optional): A config dictionary. (default: :obj:`{}`)
- **api_key** (Optional[str], optional): The API key for authenticating with the model service. (default: :obj:`None`)
- **url** (Optional[str], optional): The url to the model service. (default: :obj:`None`)
- **token_counter** (Optional[BaseTokenCounter], optional): Token counter to use for the model. If not provided, :obj:`OpenAITokenCounter` will be used. (default: :obj:`None`)
- **timeout** (Optional[float], optional): The timeout value in seconds for API calls. (default: :obj:`None`)
- **max_retries** (int, optional): Maximum number of retries for API calls. (default: :obj:`3`)
- **extract_thinking_from_response** (bool, optional): Whether to extract `__INLINE_CODE_0____INLINE_CODE_1____INLINE_CODE_2____INLINE_CODE_3____INLINE_CODE_4____INLINE_CODE_5____INLINE_CODE_6____INLINE_CODE_7____INLINE_CODE_8____INLINE_CODE_9____INLINE_CODE_10____INLINE_CODE_11__True`)
<a id="camel.models.base_model.BaseModelBackend.__init__"></a>
### __init__
```python
def __init__(
self,
model_type: Union[ModelType, str],
model_config_dict: Optional[Dict[str, Any]] = None,
api_key: Optional[str] = None,
url: Optional[str] = None,
token_counter: Optional[BaseTokenCounter] = None,
timeout: Optional[float] = Constants.TIMEOUT_THRESHOLD,
max_retries: int = 3,
extract_thinking_from_response: bool = True
):
```
<a id="camel.models.base_model.BaseModelBackend.token_counter"></a>
### token_counter
```python
def token_counter(self):
```
**Returns:**
BaseTokenCounter: The token counter following the model's
tokenization style.
<a id="camel.models.base_model.BaseModelBackend.supports_tool_response_format"></a>
### supports_tool_response_format
```python
def supports_tool_response_format(self):
```
**Returns:**
bool: :obj:`True` when the backend can natively handle
`__INLINE_CODE_1__` together with tools that are not marked
strict, otherwise :obj:`False`.
<a id="camel.models.base_model.BaseModelBackend._prepare_request_config"></a>
### _prepare_request_config
```python
def _prepare_request_config(self, tools: Optional[List[Dict[str, Any]]] = None):
```
Prepare the request configuration dictionary.
Creates a deep copy of the model config and handles tool-related
parameters. If no tools are specified, removes parallel_tool_calls
as OpenAI API only allows it when tools are present.
**Parameters:**
- **tools** (Optional[List[Dict[str, Any]]]): The tools to include in the request. (default: :obj:`None`)
**Returns:**
Dict[str, Any]: The prepared request configuration.
<a id="camel.models.base_model.BaseModelBackend._resolve_tools"></a>
### _resolve_tools
```python
def _resolve_tools(self, tools: Optional[List[Dict[str, Any]]]):
```
Resolve tools for the current request.
<a id="camel.models.base_model.BaseModelBackend._normalize_client_call_kwargs"></a>
### _normalize_client_call_kwargs
```python
def _normalize_client_call_kwargs(
self,
call: Any,
args: Any,
kwargs: Dict[str, Any]
):
```
Normalize client call arguments into a keyword dictionary.
<a id="camel.models.base_model.BaseModelBackend._mark_request_log_client_call_seen"></a>
### _mark_request_log_client_call_seen
```python
def _mark_request_log_client_call_seen(self):
```
Mark that a model client call happened during this run.
<a id="camel.models.base_model.BaseModelBackend._call_client"></a>
### _call_client
```python
def _call_client(
self,
call: Any,
*args: Any,
**kwargs: Any
):
```
Invoke a sync client call and sync request logs with payload.
<a id="camel.models.base_model.BaseModelBackend._should_sync_request_log"></a>
### _should_sync_request_log
```python
def _should_sync_request_log(self, kwargs: Dict[str, Any]):
```
Check whether a client call should sync request logs.
<a id="camel.models.base_model.BaseModelBackend._extract_model_config_from_client_kwargs"></a>
### _extract_model_config_from_client_kwargs
```python
def _extract_model_config_from_client_kwargs(self, kwargs: Dict[str, Any]):
```
Extract model config payload from client call kwargs.
<a id="camel.models.base_model.BaseModelBackend._sync_request_log_with_client_kwargs"></a>
### _sync_request_log_with_client_kwargs
```python
def _sync_request_log_with_client_kwargs(self, kwargs: Dict[str, Any]):
```
Sync request log with final payload passed to the model client.
<a id="camel.models.base_model.BaseModelBackend.preprocess_messages"></a>
### preprocess_messages
```python
def preprocess_messages(self, messages: List[OpenAIMessage]):
```
Preprocess messages before sending to model API.
Removes thinking content from assistant and user messages.
Automatically formats messages for parallel tool calls if tools are
detected.
**Parameters:**
- **messages** (List[OpenAIMessage]): Original messages.
**Returns:**
List[OpenAIMessage]: Preprocessed messages
<a id="camel.models.base_model.BaseModelBackend.postprocess_response"></a>
### postprocess_response
```python
def postprocess_response(self, response: Any):
```
Postprocess model response to extract `__INLINE_CODE_0__` tags.
For non-streaming responses, if the model embeds reasoning in
`__INLINE_CODE_1__` tags within the response content (and
`__INLINE_CODE_2__` is not already set), extracts the content
of the tags into `__INLINE_CODE_3__` and removes them from
`__INLINE_CODE_4__`.
Streaming responses and other non-ChatCompletion types are
returned as-is.
**Parameters:**
- **response**: The model response. May be a :obj:`ChatCompletion`, a streaming object, or other types.
**Returns:**
The response, potentially with `__INLINE_CODE_0__`
populated and `__INLINE_CODE_1__` cleaned.
<a id="camel.models.base_model.BaseModelBackend._log_request"></a>
### _log_request
```python
def _log_request(
self,
messages: List[OpenAIMessage],
model_config_dict: Optional[Dict[str, Any]] = None
):
```
Log the request messages to a JSON file if logging is enabled.
**Parameters:**
- **messages** (List[OpenAIMessage]): The messages to log.
- **model_config_dict** (Optional[Dict[str, Any]]): The effective model configuration used for this request.
**Returns:**
Optional[str]: The path to the log file if logging is enabled,
None otherwise.
<a id="camel.models.base_model.BaseModelBackend._log_response"></a>
### _log_response
```python
def _log_response(self, log_path: str, response: Any):
```
Log the response to the existing log file.
**Parameters:**
- **log_path** (str): The path to the log file.
- **response** (Any): The response to log.
<a id="camel.models.base_model.BaseModelBackend._log_and_trace"></a>
### _log_and_trace
```python
def _log_and_trace(self):
```
Update Langfuse trace with session metadata.
This method updates the current Langfuse trace with agent session
information and model metadata. Called at the start of _run() and
_arun() methods before API execution.
<a id="camel.models.base_model.BaseModelBackend._run"></a>
### _run
```python
def _run(
self,
messages: List[OpenAIMessage],
response_format: Optional[Type[BaseModel]] = None,
tools: Optional[List[Dict[str, Any]]] = None
):
```
Runs the query to the backend model in a non-stream mode.
**Parameters:**
- **messages** (List[OpenAIMessage]): Message list with the chat history in OpenAI API format.
- **response_format** (Optional[Type[BaseModel]]): The format of the response.
- **tools** (Optional[List[Dict[str, Any]]]): The schema of the tools to use for the request.
**Returns:**
Union[ChatCompletion, Stream[ChatCompletionChunk], Any]:
`ChatCompletion` in the non-stream mode, or
`Stream[ChatCompletionChunk]` in the stream mode,
or `ChatCompletionStreamManager[BaseModel]` in the structured
stream mode.
<a id="camel.models.base_model.BaseModelBackend.run"></a>
### run
```python
def run(
self,
messages: List[OpenAIMessage],
response_format: Optional[Type[BaseModel]] = None,
tools: Optional[List[Dict[str, Any]]] = None
):
```
Runs the query to the backend model.
**Parameters:**
- **messages** (List[OpenAIMessage]): Message list with the chat history in OpenAI API format.
- **response_format** (Optional[Type[BaseModel]]): The response format to use for the model. (default: :obj:`None`)
- **tools** (Optional[List[Tool]]): The schema of tools to use for the model for this request. Will override the tools specified in the model configuration (but not change the configuration). (default: :obj:`None`)
**Returns:**
Union[ChatCompletion, Stream[ChatCompletionChunk], Any]:
`ChatCompletion` in the non-stream mode,
`Stream[ChatCompletionChunk]` in the stream mode, or
`ChatCompletionStreamManager[BaseModel]` in the structured
stream mode.
<a id="camel.models.base_model.BaseModelBackend.count_tokens_from_messages"></a>
### count_tokens_from_messages
```python
def count_tokens_from_messages(self, messages: List[OpenAIMessage]):
```
Count the number of tokens in the messages using the specific
tokenizer.
**Parameters:**
- **messages** (List[Dict]): message list with the chat history in OpenAI API format.
**Returns:**
int: Number of tokens in the messages.
<a id="camel.models.base_model.BaseModelBackend._to_chat_completion"></a>
### _to_chat_completion
```python
def _to_chat_completion(self, response: ParsedChatCompletion):
```
<a id="camel.models.base_model.BaseModelBackend.token_limit"></a>
### token_limit
```python
def token_limit(self):
```
**Returns:**
int: The maximum token limit for the given model.
<a id="camel.models.base_model.BaseModelBackend.stream"></a>
### stream
```python
def stream(self):
```
**Returns:**
bool: Whether the model is in stream mode.