ahvn.agent.conv_agent module

class ahvn.agent.conv_agent.ConvToolAgentSpec(prompt, tools=None, llm_args=None, max_steps=None, **kwargs)

Bases: BasePromptAgentSpec

Warning: ConvToolAgentSpec is only used to accomodate LLMs that do not support native function calls. Whenever possible, for best performance and stability, it is recommended to use LLM.tooluse, LLM.oracle(tools=…), SubmitToolAgentSpec, or other tool use agents instead.

Parameters:

• prompt (PromptUKFT)

• tools (List[ToolSpec] | None)

• llm_args (Dict | None)

• max_steps (int | None)

__init__(prompt, tools=None, llm_args=None, max_steps=None, **kwargs)

Parameters:

• prompt (PromptUKFT)

• tools (List[ToolSpec] | None)

• llm_args (Dict | None)

• max_steps (int | None)

is_done(messages, delta_messages)

Return type:

Tuple[bool, Dict[str, Any]]

Parameters:

• messages (str | Dict[str, Any] | Any | List[str | Dict[str, Any] | Any])

• delta_messages (List[Dict[str, Any]])

decode(messages, finish_state=None)

Return type:

Any

Parameters:

• messages (str | Dict[str, Any] | Any | List[str | Dict[str, Any] | Any])

• finish_state (Dict[str, Any])

step(messages, include=None)

Execute a single step with conversational tool calling.

This step streams LLM output, then parses for tool calls and executes them.

Return type:

Generator[Dict[str, Any], None, None]

Parameters:

• messages (str | Dict[str, Any] | Any | List[str | Dict[str, Any] | Any])

• include (List[Literal['text', 'think', 'tool_calls', 'content', 'message', 'structured', 'tool_messages', 'tool_results', 'delta_messages', 'messages']] | None)