automl

func DeployModel ¶

func DeployModel(name string) (*automl.DeployModelOperation, error)

DeployModel deploys the specified model

func GetLatestDeployed ¶

func GetLatestDeployed(projectID string, location string, modelName string) (*automlpb.Model, error)

GetLatestDeployed finds the latest deployed model

TODO(jlewi): We should really filter on labels; they don't appear to show up in the UI but they are in the API: https://cloud.google.com/automl/docs/reference/rest/v1/projects.locations.models#Model

func GetLatestTrained ¶

func GetLatestTrained(projectID string, location string, modelName string) (*automlpb.Model, error)

GetLatestTrained finds the latest trained model

TODO(jlewi): We should really filter on labels; they don't appear to show up in the UI but they are in the API: https://cloud.google.com/automl/docs/reference/rest/v1/projects.locations.models#Model

func GetModel ¶

func GetModel(name string) (*automlpb.Model, error)

GetModel gets the specified model

func GetModelEvaluation ¶

func GetModelEvaluation(name string, outputFile string) (*automlpb.ModelEvaluation, error)

GetModelEvaluation gets the evaluation for the specified model

func GetOperation ¶

func GetOperation(name string) (*lropb.Operation, error)

GetOperation gets the specified operation

func IsTraining ¶

func IsTraining(projectID string, location string) (bool, error)

IsTraining determines if there is an import job or model training currently in progress

TODO(jlewi): We should really only look at models and datasets matching some labels or other criterion. Its not clear how to link LRO's for import jobs to the corresponding dataset. So right now we just check if there are any operations still in progress.

func LabelModel ¶

func LabelModel(name string, labels map[string]string) error

labelModel labels the specified model

TODO(jlewi): Should we support removing labels?