KitaruSyncClient

Synchronous Kitaru client.

Attributes

attributeapiKitaruAPIClient

API client with one method per endpoint.

Functions

constructor__init__(api_client=None) -> None

Initialize the client.

paramapi_clientKitaruAPIClient | None

= None

API client used to send requests.

Returns

None

funcclose() -> None

Close the underlying client and stop the event loop thread.

Returns

None

funcget_agent(agent) -> AgentResponse

Get an agent by id or name.

paramagentuuid.UUID | str

Id or name of the agent.

Returns

kitaru.api_models.v1.agent.AgentResponse

Stored agent.

funcregister_agent

(name, run_spec, *, description=None, display_version=None, version_description=None, capabilities=None, agent_idempotency_key=None, version_idempotency_key=None) -> AgentRegistrationResult

Create an agent and its initial version.

paramnamestr

New agent name.

paramrun_specRunSpec

Run spec for the initial version.

paramdescriptionstr | None

= None

Agent description.

paramdisplay_versionstr | None

= None

Human-readable designator for the initial version.

paramversion_descriptionstr | None

= None

Initial version description.

paramcapabilitiesAgentCapabilities | None

= None

Initial version capabilities.

paramagent_idempotency_keystr | None

= None

Idempotency key for agent creation.

paramversion_idempotency_keystr | None

= None

Idempotency key for version creation.

Returns

kitaru.client.AgentRegistrationResult

Created agent and initial version.

funcregister_agent_version

(agent, run_spec, *, display_version=None, description=None, capabilities=None, idempotency_key=None) -> AgentVersionResponse

Create the next version of an existing agent.

paramagentuuid.UUID | str

Id or exact name of the agent.

paramrun_specRunSpec

Run spec for the new version.

paramdisplay_versionstr | None

= None

Human-readable version designator.

paramdescriptionstr | None

= None

Version description.

paramcapabilitiesAgentCapabilities | None

= None

Version capabilities.

paramidempotency_keystr | None

= None

Idempotency key for version creation.

Returns

kitaru.api_models.v1.agent_version.AgentVersionResponse

Created agent version.

funclist_agents() -> Iterator[AgentResponse]

Iterate over all agents.

Returns

collections.abc.Iterator

Iterator over every agent.

funcget_session(session_id) -> SessionDetailResponse

Get a session by id.

paramsession_iduuid.UUID

Id of the session.

Returns

kitaru.api_models.v1.session.SessionDetailResponse

Stored session.

funclist_sessions(agent=None) -> Iterator[SessionResponse]

Iterate over sessions, optionally scoped to one agent.

paramagentuuid.UUID | str | None

= None

Id or name of the agent to scope to.

Returns

collections.abc.Iterator

Iterator over every matching session.

funclist_session_nodes(session_id) -> Iterator[SessionNodeResponse]

Iterate over the nodes of a session in index order.

paramsession_iduuid.UUID

Id of the session.

Returns

collections.abc.Iterator

Iterator over every node.

funcreplay

(session_id, evaluators, agent_version_id=None, override=None, tool_policy=None, baseline_evaluation_mode=BaselineEvaluationMode.IF_MISSING, wait=True, timeout=None) -> ReplayResponse

Replay a session and optionally wait for it to finish.

paramsession_iduuid.UUID

Id of the session to replay.

paramevaluatorslist[EvaluatorConfig]

Evaluators run against the result session.

paramagent_version_iduuid.UUID | None

= None

Agent version to replay with, the session's recorded version when unset.

paramoverrideReplayOverride | None

= None

Override to apply.

paramtool_policyToolPolicy | None

= None

Tool policy to apply.

parambaseline_evaluation_modeBaselineEvaluationMode

= BaselineEvaluationMode.IF_MISSING

How to score the baseline session.

paramwaitbool

= True

Whether to wait for the replay to reach a terminal status.

paramtimeoutfloat | None

= None

Seconds to wait before giving up.

Returns

kitaru.api_models.v1.replay.ReplayResponse

Replay, in a terminal status when waited for.

funcget_replay(replay_id) -> ReplayResponse

Get a replay by id.

paramreplay_iduuid.UUID

Id of the replay.

Returns

kitaru.api_models.v1.replay.ReplayResponse

Stored replay.

funcwait_for_replay(replay_id, timeout=None, poll_interval=2.0) -> ReplayResponse

Wait for a replay to reach a terminal status.

paramreplay_iduuid.UUID

Id of the replay.

paramtimeoutfloat | None

= None

Seconds to wait before giving up.

parampoll_intervalfloat

= 2.0

Seconds between status checks.

Returns

kitaru.api_models.v1.replay.ReplayResponse

Replay in a terminal status.

funcget_experiment(experiment) -> ExperimentResponse

Get an experiment by id or name.

paramexperimentuuid.UUID | str

Id or name of the experiment.

Returns

kitaru.api_models.v1.experiment.ExperimentResponse

Stored experiment.

funclist_experiments() -> Iterator[ExperimentResponse]

Iterate over all experiments.

Returns

collections.abc.Iterator

Iterator over every experiment.

funcrun_experiment

(experiment, cohort_version_id, agent_version_id, baseline_evaluation_mode=BaselineEvaluationMode.IF_MISSING, wait=True, timeout=None) -> ExperimentRunResponse

Start an experiment run and optionally wait for it to finish.

paramexperimentuuid.UUID | str

Id or name of the experiment.

paramcohort_version_iduuid.UUID

Cohort version whose sessions are replayed.

paramagent_version_iduuid.UUID

Agent version to replay with.

parambaseline_evaluation_modeBaselineEvaluationMode

= BaselineEvaluationMode.IF_MISSING

How to score each baseline session.

paramwaitbool

= True

Whether to wait for the run to reach a terminal status.

paramtimeoutfloat | None

= None

Seconds to wait before giving up.

Returns

kitaru.api_models.v1.experiment_run.ExperimentRunResponse

Experiment run, in a terminal status when waited for.

funcget_experiment_run(run_id) -> ExperimentRunResponse

Get an experiment run by id.

paramrun_iduuid.UUID

Id of the experiment run.

Returns

kitaru.api_models.v1.experiment_run.ExperimentRunResponse

Stored experiment run.

funcwait_for_experiment_run(run_id, timeout=None, poll_interval=2.0) -> ExperimentRunResponse

Wait for an experiment run to reach a terminal status.

paramrun_iduuid.UUID

Id of the experiment run.

paramtimeoutfloat | None

= None

Seconds to wait before giving up.

parampoll_intervalfloat

= 2.0

Seconds between status checks.

Returns

kitaru.api_models.v1.experiment_run.ExperimentRunResponse

Experiment run in a terminal status.

AttributesCopy Anchor Link

FunctionsCopy Anchor Link

On this page

Attributes

Functions