Model Groups

class indico.queries.model_groups.AddModelGroupComponent(workflow_id, dataset_id, name, source_column_id, after_component_id=None, after_link_id=None, labelset_column_id=None, new_labelset_args=None, new_questionnaire_args=None, model_training_options=None, model_type=None)

Adds a new model group to a workflow, optionally with a customized questionnaire.
Available on 5.0+ only.
Returns workflow with updated component list, which will contain the added Model Group.

  • Parameters:
    • workflow_id (int) – the id of the workflow to add the component to.
    • dataset_id (int) – the id of the dataset of this workflow/model group.
    • name (str) – name of the model group.
    • source_column_id (str) – source column identifier from dataset.
    • after_component_id (str) – the id of the previous component, or step, that should precede this step.
      Typically, this prior component would be something like “INPUT_OCR_EXTRACTION” or “INPUT_IMAGE” and can
      be found by using “component_by_type()” on the parent workflow.
    • labelset_column_id (int) – the labelset column to copy from.
    • new_labelset_args (NewLabelsetArguments) – if needed, new labelset to add.
      Only use if not using labelset_column_id.
    • new_questionnaire_args (NewQuestionnaireArguments) – Customize the questionnaire associated with this model group.

class indico.queries.model_groups.ModelGroupPredict(model_id, data, load=True, predict_options=None)

Generate predictions from a model group on new data

  • Parameters:
    • model_id (int) – selected model id use for predictions
    • data (List[str**]) – list of samples to predict
    • predict_options (JSONString) – arguments for predictions
  • Returns:
    Job associated with this model group predict task

class indico.queries.model_groups.GetModelGroup(id, wait=False)

Get an object describing a model group

  • Parameters:
    id (int) – model group id to query
  • Returns:
    ModelGroup object

class indico.queries.model_groups.GetModelGroupSelectedModelStatus(id)

Get the status string of the selected model for the given model group id

  • Parameters:
    id (int) – model group id to query
  • Returns:
    CREATED, TRAINING, COMPLETE or FAILED
  • Return type:
    status (str)