Function functionary_chat_handler

llama_cpp/llama_chat_format.py:1461–1819

(
    llama: llama.Llama,
    messages: List[llama_types.ChatCompletionRequestMessage],
    functions: Optional[List[llama_types.ChatCompletionFunction]] = None,
    function_call: Optional[llama_types.ChatCompletionRequestFunctionCall] = None,
    tools: Optional[List[llama_types.ChatCompletionTool]] = None,
    tool_choice: Optional[llama_types.ChatCompletionToolChoiceOption] = None,
    temperature: float = 0.2,
    top_p: float = 0.95,
    top_k: int = 40,
    min_p: float = 0.05,
    typical_p: float = 1.0,
    stream: bool = False,
    stop: Optional[Union[str, List[str]]] = [],
    response_format: Optional[llama_types.ChatCompletionRequestResponseFormat] = None,
    max_tokens: Optional[int] = None,
    presence_penalty: float = 0.0,
    frequency_penalty: float = 0.0,
    repeat_penalty: float = 1.1,
    tfs_z: float = 1.0,
    mirostat_mode: int = 0,
    mirostat_tau: float = 5.0,
    mirostat_eta: float = 0.1,
    model: Optional[str] = None,
    logits_processor: Optional[llama.LogitsProcessorList] = None,
    grammar: Optional[llama.LlamaGrammar] = None,
    **kwargs,  # type: ignore
)

@register_chat_completion_handler("functionary")
def functionary_chat_handler(
    llama: llama.Llama,
    messages: List[llama_types.ChatCompletionRequestMessage],
    functions: Optional[List[llama_types.ChatCompletionFunction]] = None,
    function_call: Optional[llama_types.ChatCompletionRequestFunctionCall] = None,
    tools: Optional[List[llama_types.ChatCompletionTool]] = None,
    tool_choice: Optional[llama_types.ChatCompletionToolChoiceOption] = None,
    temperature: float = 0.2,
    top_p: float = 0.95,
    top_k: int = 40,
    min_p: float = 0.05,
    typical_p: float = 1.0,
    stream: bool = False,
    stop: Optional[Union[str, List[str]]] = [],
    response_format: Optional[llama_types.ChatCompletionRequestResponseFormat] = None,
    max_tokens: Optional[int] = None,
    presence_penalty: float = 0.0,
    frequency_penalty: float = 0.0,
    repeat_penalty: float = 1.1,
    tfs_z: float = 1.0,
    mirostat_mode: int = 0,
    mirostat_tau: float = 5.0,
    mirostat_eta: float = 0.1,
    model: Optional[str] = None,
    logits_processor: Optional[llama.LogitsProcessorList] = None,
    grammar: Optional[llama.LlamaGrammar] = None,
    **kwargs,  # type: ignore
) -> Union[llama_types.ChatCompletion, Iterator[llama_types.ChatCompletionChunk]]:
    SYSTEM_MESSAGE = """A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. The assistant calls functions with appropriate input when necessary"""

    def generate_type_definition(
        param: Dict[str, llama_types.JsonType], indent_level: int, shared_defs
    ) -> str:
        indent = "  " * indent_level
        if "$ref" in param:
            # Reference to a shared definition
            ref_name = param["$ref"].split("/")[
                -1
            ]  # Extract the type name from the reference
            return ref_name
        elif param.get("type") == "array":
            items = param.get("items", {})
            item_type = generate_type_definition(items, indent_level + 1, shared_defs)
            return f"Array<{item_type}>"
        elif param.get("type") == "object":
            properties = param.get("properties", {})
            nested_schema = "{\n"
            for nested_param_name, nested_param in properties.items():
                nested_param_type = generate_type_definition(
                    nested_param, indent_level + 1, shared_defs
                )
                nested_schema += (
                    f"{indent}  {nested_param_name}: {nested_param_type},\n"
                )
            nested_schema += indent + "}"
            return nested_schema
        elif "enum" in param:
            # Enum type