prebuilt.gtla.models¶

Classes¶

`AgentAction`	Model representing an agent's action decision.
`AgentOutput`	Model representing the output from an agent's action.

Module Contents¶

class prebuilt.gtla.models.AgentAction(/, **data: Any)¶

Bases: pydantic.BaseModel

Model representing an agent’s action decision.

action¶

The specific action to be taken (e.g., “search_calendar”, “analyze_tasks”)

Type:: str

thought¶

The reasoning process behind the action choice

Type:: str

tool¶

The specific tool to be used for the action (if needed)

Type:: Optional[str]

action_input¶

Input parameters for the action

Type:: Optional[Dict]

Examples

>>> action = AgentAction(
...     action="search_calendar",
...     thought="Need to check schedule conflicts",
...     tool="calendar_search",
...     action_input={"date_range": "next_week"}
... )

action: str¶

action_input: dict | None = None¶

thought: str¶

tool: str | None = None¶

class prebuilt.gtla.models.AgentOutput(/, **data: Any)¶

Bases: pydantic.BaseModel

Model representing the output from an agent’s action.

observation¶

The result or observation from executing the action

Type:: str

output¶

Structured output data from the action

Type:: Dict

Examples

>>> output = AgentOutput(
...     observation="Found 3 free time slots next week",
...     output={
...         "free_slots": ["Mon 2PM", "Wed 10AM", "Fri 3PM"],
...         "conflicts": []
...     }
... )

observation: str¶

output: dict¶