Skip to content

orca_sdk.client#

orca_api module-attribute #

orca_api = OrcaClient(
    event_hooks={
        "request": [_instrument_request],
        "response": [_raise_error_for_response],
    },
    follow_redirects=True,
    timeout=Timeout(connect=3, read=20, write=10, pool=5),
)

Typed client for the Orca API

ActionRecommendation #

Bases: TypedDict

action instance-attribute #

action

The recommended action to take

rationale instance-attribute #

rationale

Explanation for why this action was recommended

PredictionFeedbackRequest #

Bases: TypedDict

value instance-attribute #

value

The feedback value. For updates, UNSET means keep existing value. None means delete the feedback.

comment instance-attribute #

comment

Optional comment. For updates, UNSET means keep existing comment. None means remove the comment.

GetMemorysetByNameOrIdMemoryByMemoryIdParams #

Bases: TypedDict

memory_id instance-attribute #

memory_id

ID of the memory

DeleteMemorysetByNameOrIdMemoryByMemoryIdParams #

Bases: TypedDict

memory_id instance-attribute #

memory_id

ID of the memory

PostDatasourceUploadRequest #

Bases: TypedDict

name instance-attribute #

name

Name for the datasource

description instance-attribute #

description

Optional description for the datasource

GetDatasourceByNameOrIdDownloadParams #

Bases: TypedDict

file_type instance-attribute #

file_type

File type to download: * hf_dataset: Zipped HuggingFace dataset (default) * json: Row-oriented JSON array * csv: CSV file

GetTelemetryPredictionByPredictionIdMemorySuggestionsParams #

Bases: TypedDict

prediction_id instance-attribute #

prediction_id

ID of the prediction to generate suggestions for

num_memories instance-attribute #

num_memories

Number of memory suggestions to generate

refresh instance-attribute #

refresh

Force the explanation agent to re-run even if a cached explanation exists