Skip to content

armgit / camel

Public Imported from GitHub
0

🐫 CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents. https://www.camel-ai.org

6.6 KB
Raw Blame
<a id="camel.models.deepseek_model"></a>

<a id="camel.models.deepseek_model.DeepSeekModel"></a>

## DeepSeekModel

```python
class DeepSeekModel(OpenAICompatibleModel):
```

DeepSeek API in a unified OpenAICompatibleModel interface.

**Parameters:**

- **model_type** (Union[ModelType, str]): Model for which a backend is created.
- **model_config_dict** (Optional[Dict[str, Any]], optional): A dictionary that will be fed into:obj:`openai.ChatCompletion.create()`. If :obj:`None`, :obj:`DeepSeekConfig().as_dict()` will be used. (default: :obj:`None`)
- **api_key** (Optional[str], optional): The API key for authenticating with the DeepSeek service. (default: :obj:`None`)
- **url** (Optional[str], optional): The url to the DeepSeek service. (default: :obj:`https://api.deepseek.com`)
- **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. If not provided, will fall back to the MODEL_TIMEOUT environment variable or default to 180 seconds. (default: :obj:`None`)
- **max_retries** (int, optional): Maximum number of retries for API calls. (default: :obj:`3`) **kwargs (Any): Additional arguments to pass to the client initialization.
- **References**: 
- **https**: //api-docs.deepseek.com/

<a id="camel.models.deepseek_model.DeepSeekModel.__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] = None,
    max_retries: int = 3,
    **kwargs: Any
):
```

<a id="camel.models.deepseek_model.DeepSeekModel._get_reasoning_session_id"></a>

### _get_reasoning_session_id

```python
def _get_reasoning_session_id(self):
```

Get the current agent session ID for reasoning cache isolation.

<a id="camel.models.deepseek_model.DeepSeekModel._get_tool_call_id"></a>

### _get_tool_call_id

```python
def _get_tool_call_id(tool_call: Any):
```

<a id="camel.models.deepseek_model.DeepSeekModel._get_message_content"></a>

### _get_message_content

```python
def _get_message_content(message: Any):
```

<a id="camel.models.deepseek_model.DeepSeekModel._cache_tool_call_reasoning"></a>

### _cache_tool_call_reasoning

```python
def _cache_tool_call_reasoning(self, response: Any):
```

Cache reasoning content for DeepSeek thinking-mode tool calls.

DeepSeek requires the assistant `__INLINE_CODE_0__` generated with a
tool call to be passed back in later requests. CAMEL memory stores the
OpenAI-compatible tool-call message, so the model layer keeps the
DeepSeek-specific field keyed by tool_call_id and injects it before
the next request.

<a id="camel.models.deepseek_model.DeepSeekModel._cache_reasoning_content"></a>

### _cache_reasoning_content

```python
def _cache_reasoning_content(
    self,
    reasoning_content: Optional[str],
    tool_call_ids: List[str],
    content: Optional[str] = None
):
```

Cache DeepSeek reasoning collected from a streamed response.

<a id="camel.models.deepseek_model.DeepSeekModel._collect_stream_reasoning"></a>

### _collect_stream_reasoning

```python
def _collect_stream_reasoning(
    self,
    chunk: ChatCompletionChunk,
    reasoning_parts: List[str],
    content_parts: List[str],
    tool_call_ids: Set[str]
):
```

Collect reasoning/content/tool ids from a streaming chunk.

<a id="camel.models.deepseek_model.DeepSeekModel._is_stream_response_complete"></a>

### _is_stream_response_complete

```python
def _is_stream_response_complete(chunk: ChatCompletionChunk):
```

Return whether a streaming chunk completes one response.

<a id="camel.models.deepseek_model.DeepSeekModel._cache_collected_stream_reasoning"></a>

### _cache_collected_stream_reasoning

```python
def _cache_collected_stream_reasoning(
    self,
    reasoning_parts: List[str],
    content_parts: List[str],
    tool_call_ids: Set[str]
):
```

Cache collected streaming reasoning once it is complete.

<a id="camel.models.deepseek_model.DeepSeekModel._wrap_stream_reasoning_cache"></a>

### _wrap_stream_reasoning_cache

```python
def _wrap_stream_reasoning_cache(self, response: Stream[ChatCompletionChunk]):
```

Wrap a DeepSeek stream and cache reasoning after consumption.

<a id="camel.models.deepseek_model.DeepSeekModel._inject_tool_call_reasoning"></a>

### _inject_tool_call_reasoning

```python
def _inject_tool_call_reasoning(self, messages: List[OpenAIMessage]):
```

Inject cached reasoning into assistant tool-call messages.

<a id="camel.models.deepseek_model.DeepSeekModel._with_reasoning_content"></a>

### _with_reasoning_content

```python
def _with_reasoning_content(self, message: OpenAIMessage, reasoning_content: str):
```

Return a copy of an OpenAI message with DeepSeek reasoning.

<a id="camel.models.deepseek_model.DeepSeekModel.preprocess_messages"></a>

### preprocess_messages

```python
def preprocess_messages(self, messages: List[OpenAIMessage]):
```

Preprocess messages and restore DeepSeek tool-call reasoning.

<a id="camel.models.deepseek_model.DeepSeekModel.postprocess_response"></a>

### postprocess_response

```python
def postprocess_response(self, response: Any):
```

Postprocess responses and cache DeepSeek tool-call reasoning.

<a id="camel.models.deepseek_model.DeepSeekModel._normalize_thinking_config"></a>

### _normalize_thinking_config

```python
def _normalize_thinking_config(self, request_config: Dict[str, Any]):
```

Move top-level `__INLINE_CODE_0____INLINE_CODE_1____INLINE_CODE_2__` for OpenAI SDK.

<a id="camel.models.deepseek_model.DeepSeekModel._is_thinking_mode_enabled"></a>

### _is_thinking_mode_enabled

```python
def _is_thinking_mode_enabled(self, request_config: Dict[str, Any]):
```

<a id="camel.models.deepseek_model.DeepSeekModel._prepare_request"></a>

### _prepare_request

```python
def _prepare_request(
    self,
    messages: List[OpenAIMessage],
    response_format: Optional[Type[BaseModel]] = None,
    tools: Optional[List[Dict[str, Any]]] = None
):
```

<a id="camel.models.deepseek_model.DeepSeekModel._run"></a>

### _run

```python
def _run(
    self,
    messages: List[OpenAIMessage],
    response_format: Optional[Type[BaseModel]] = None,
    tools: Optional[List[Dict[str, Any]]] = None
):
```

Runs inference of DeepSeek chat completion.

**Parameters:**

- **messages** (List[OpenAIMessage]): Message list with the chat history in OpenAI API format.

**Returns:**

  Union[ChatCompletion, Stream[ChatCompletionChunk]]:
`ChatCompletion` in the non-stream mode, or
`Stream[ChatCompletionChunk]` in the stream mode.