Model store APIs
Operations for creating and managing models, model tags, policies, segments, and model alerts. This data is maintained in the VIANOPS model store database.
vianops_client.api.modelstore.alerts
- class AlertsV1API(config: Config | None = None)
Bases:
ConfiguredBaseApiClass
- create(data: V1AlertCreateModelList, output_type: str = 'object') V1AlertModelList
Create alerts
- Parameters:
data (V1AlertCreateModelList) – An instance of V1AlertCreateModelList containing the data for the new alerts.
- Returns:
An instance of V1AlertModelList containing information about the created alerts.
- Return type:
- delete(delete_params: V1AlertSearchModel, deletetype: str = 'archive', output_type: str = 'object') VianaiSuccessResponse
Delete or archive alerts
- Parameters:
delete_params (V1AlertSearchModel) – Instance containing the required data to select, and archive or delete selected alerts.
deletetype (str) – The deletion type. Supported values: archive, delete, force_delete. Default: archive
- Returns:
Instance containing response from alert deletion.
- Return type:
- delete_url = '/v1/monitor-alerts/{deletetype}'
- search(search_params: V1AlertSearchRequest, output_type: str = 'object') V1AlertPageModel
Search for alerts
- Parameters:
search_params (V1AlertSearchRequest) – An instance of V1AlertSearchRequest containing the search parameters for alerts.
- Returns:
An instance of V1AlertPageModel containing information about the searched alerts.
- Return type:
- search_url = '/v1/monitor-alerts/search'
- update(update_params: V1AlertUpdateModelList, output_type: str = 'object') V1AlertUpdateResponseModel
Update existing alerts
- Parameters:
update_params (V1AlertUpdateModelList) – An instance of V1AlertUpdateModelList containing the data for the alert updates.
- Returns:
An instance of V1AlertUpdateResponseModel containing information about the alert updates.
- Return type:
- url_suffix = '/v1/monitor-alerts'
vianops_client.api.modelstore.feature_importance_file_utils
- check_absolute_percentage_and_sum(df)
- check_and_convert_file_to_json(content, name, tag_name)
vianops_client.api.modelstore.models
- class ModelsV1Api(config: Config | None = None)
Bases:
ConfiguredBaseApiClass
- create(data: V1ModelsModelList, output_type: str = 'object') V1ModelsModelList
Create models
- Parameters:
data (V1ModelsModelList) – Instance containing the data to create models.
- Returns:
Instance containing a list of the created models.
- Return type:
- delete(data: V1ModelsFilters, deletetype: str = 'archive', output_type: str = 'object') V1ModelsModelList
Delete or archive models
- Parameters:
data (V1ModelsFilters) – Instance containing the required data to select, and archive or delete selected models.
deletetype (str) – The deletion type. Supported values: archive, delete, force_delete. Default: archive
- Returns:
Instance containing response from model deletion.
- Return type:
- delete_url = '/v1/models/{deletetype}'
- search(data: V1ModelSearch, output_type: str = 'object') V1ModelsModelPage
Search for models
- Parameters:
data (V1ModelSearch) – Instance containing data to be use to perform a paginated model search.
- Returns:
Instance containing a list of models collected from search along with pagination details.
- Return type:
- search_url = '/v1/models/search'
- update(data: V1ModelUpdates, output_type: str = 'object') V1ModelsModelList
Update existing models
- Parameters:
data (V1ModelUpdates) – An instance of V1ModelUpdates containing the data for updating the model.
- Returns:
Instance containing information about the updated models.
- Return type:
- url_suffix = '/v1/models'
vianops_client.api.modelstore.policies
- class PoliciesV1Api(config: Config | None = None)
Bases:
ConfiguredBaseApiClass
- bulk_create(project_name: str = '', model_name: str = '', model_version: str | None = '1', model_stage: str | None = 'primary', window_type: List[str] | None = ['day', 'week', 'month', 'quarter'], window_method=None, features=None, name_prefix=None, hotspot_analysis: dict | None = None, segments=None, warning_level: float | None = 0.1, critical_level: float | None = 0.5, drift_measure: str | None = 'PSI', schedule: str | None = '0 0 6 ? * *', output_type: str = 'object', training_baseline=False)
- create(data: V1PolicyRequestModelList, output_type: str = 'object') V1PolicyModelList
Create policies
- Parameters:
data (V1PolicyRequestModelList) – Instance containing data to create policies.
- Returns:
Instance containing a list of the created policies.
- Return type:
- delete(data: V1PolicyModelSearch, deletetype: str = 'archive', output_type: str = 'object') VianaiSuccessResponse
Delete or archive policies
- Parameters:
data (V1PolicyModelSearch) – Instance containing the required data to select, and archive or delete the selected policies.
deletetype (str) – The deletion type. Supported values: archive, delete, force_delete. Default: archive.
- Returns:
Instance containing response from policy deletion.
- Return type:
- delete_url = '/v1/policies/{deletetype}'
- run_policies(deployments: List[str] = [], model_names: List[str] = [], model_versions: List[str] = ['1'], model_stages: List[str] = ['primary'], start_date: datetime | None = None, end_date: datetime | None = None, offset_target: List[int] = [1], offset_type: str = '', policy_names: List[str] | None = None, activate_policies: bool = True, num_threads: int = 15) List[dict]
Run policies
- Parameters:
deployments (List[str]) – A list of deployment names to filter policies by.
model_names (List[str]) – A list of model names to filter policies by. Default: value of deployments.
model_versions (List[str]) – A list of model versions to filter policies by. Default: [“1”].
model_stages (List[str]) – A list of model stages to filter policies by. Default: [“primary”].
start_date (Optional[datetime]) – First date to start running polcicies from.
end_date (Optional[datetime]) – Last date to run policies to.
offset_target (List[int]) – A list of offsets to be applied to policy run date. If value not provided and start_date and end_date are not provided, default will be used and all policies will run with an offset of 1.
offset_type (Optional[str]) – Type of offset. Supported values: day, week, month, quarter.
policy_names (Optional[List[str]]) – A list of policy names to run to filter policies by.
activate_policies (bool) – Indicates whether to activate policies or not. If True, policies will run accordingly on their specified schedule; otherwise, policies will only run once. Default: True.
num_threads (int) – The number of threads to running the policies. Default: 15.
- Returns:
A list of dictionaries containing the responses from each policy run.
- Return type:
List[dict]
- search(data: V1PolicySearchRequest, output_type: str = 'object') V1PolicyModelPage
Search for policies
- Parameters:
data (V1PolicySearchRequest) – Instance containing data to be used to perform a paginated segment search.
- Returns:
Instance containing a list of policies collected from search along with pagination details.
- Return type:
- search_url = '/v1/policies/search'
- update(data: V1PolicyUpdateModelList, output_type: str = 'object') V1PolicyModelList
Update policies
- Parameters:
data (V1PolicyUpdateModelList) – Instance containing the required data to select and update selected policies.
- Returns:
Instance containing information about the updated policies.
- Return type:
- url_suffix = '/v1/policies'
vianops_client.api.modelstore.segments
- class SegmentsV1Api(config: Config | None = None)
Bases:
ConfiguredBaseApiClass
- create(data: V1SegmentBaseModelList, output_type: str = 'object') V1SegmentModelList
Create segments
- Parameters:
data (V1SegmentBaseModelList) – Instance containing data to create segments.
- Returns:
Instance containing a list of the created segments.
- Return type:
- delete(data: V1SegmentSearch, deletetype: str = 'archive', output_type: str = 'object') VianaiSuccessResponse
Delete or archive segments
- Parameters:
data (V1SegmentSearch) – Instance containing the required data to select, and archive or delete the selected segments.
deletetype (str) – The deletion type. Supported values: archive, delete, force_delete. Default: archive.
- Returns:
Instance containing response from segment deletion.
- Return type:
- delete_url = '/v1/segments/{deletetype}'
- get_filter(data: dict, output_type: str = 'object') V1SegmentFilterCondition
Generate the conditions for a segment
- Parameters:
data (dict) – A dictionary containing the segment id, database connection, and table name to generate filter condition. Sample dictionary object:` {“id”: 1, “connection”: “connection_string”, “table”: “table_name”}`
- Returns:
Instance containing information about the condition.
- Return type:
- get_preview(data: V1SegmentPreviewRequestModel, output_type: str = 'object') V1SegmentPreviewModel
Get a preview of the segment data
- Parameters:
data (V1SegmentPreviewRequestModel) – Instance containing the required data to update segments.
- Returns:
Instance containing segment preview data
- Return type:
- preview_url = '/v1/segments/preview'
- search(data: V1SegmentPageSearch, output_type: str = 'object') V1SegmentModelPage
Search for segments
- Parameters:
data (V1SegmentPageSearch) – Instance containing data to be use to perform a paginated segment search.
- Returns:
Instance containing a list of segments collected from search along with pagination details.
- Return type:
- search_url = '/v1/segments/search'
- update(data: V1SegmentUpdateRequestModelList, output_type: str = 'object') V1SegmentModelList
Update existing segments
- Parameters:
data (V1SegmentUpdateRequestModelList) – Instance containing the required data to select and update selected segments.
- Returns:
Instance containing information about the updated segments.
- Return type:
- url_id_filter = '/v1/segments/{id}/filter?connection={connection}&table={table}'
- url_suffix = '/v1/segments'