Playbook(mapping=None, *, ignore_unknown_fields=False, **kwargs)
Playbook is the basic building block to instruct the LLM how to execute a certain task.
A playbook consists of a goal to accomplish, an optional list of step by step instructions (the step instruction may refers to name of the custom or default plugin tools to use) to perform the task, a list of contextual input data to be passed in at the beginning of the invoked, and a list of output parameters to store the playbook result.
Attributes |
|
---|---|
Name | Description |
name |
str
The unique identifier of the playbook. Format: projects/ .
|
display_name |
str
Required. The human-readable name of the playbook, unique within an agent. |
goal |
str
Required. High level description of the goal the playbook intend to accomplish. |
input_parameter_definitions |
MutableSequence[google.cloud.dialogflowcx_v3beta1.types.ParameterDefinition]
Optional. Defined structured input parameters for this playbook. |
output_parameter_definitions |
MutableSequence[google.cloud.dialogflowcx_v3beta1.types.ParameterDefinition]
Optional. Defined structured output parameters for this playbook. |
instruction |
google.cloud.dialogflowcx_v3beta1.types.Playbook.Instruction
Instruction to accomplish target goal. |
token_count |
int
Output only. Estimated number of tokes current playbook takes when sent to the LLM. |
create_time |
google.protobuf.timestamp_pb2.Timestamp
Output only. The timestamp of initial playbook creation. |
update_time |
google.protobuf.timestamp_pb2.Timestamp
Output only. Last time the playbook version was updated. |
referenced_playbooks |
MutableSequence[str]
Output only. The resource name of other playbooks referenced by the current playbook in the instructions. |
referenced_flows |
MutableSequence[str]
Output only. The resource name of flows referenced by the current playbook in the instructions. |
referenced_tools |
MutableSequence[str]
Optional. The resource name of tools referenced by the current playbook in the instructions. If not provided explicitly, they are will be implied using the tool being referenced in goal and steps. |
llm_model_settings |
google.cloud.dialogflowcx_v3beta1.types.LlmModelSettings
Optional. Llm model settings for the playbook. |
Classes
Instruction
Instruction(mapping=None, *, ignore_unknown_fields=False, **kwargs)
Message of the Instruction of the playbook.
Step
Step(mapping=None, *, ignore_unknown_fields=False, **kwargs)
Message of single step execution.
.. _oneof: https://proto-plus-python.readthedocs.io/en/stable/fields.html#oneofs-mutually-exclusive-fields