Export Model to ModelOps

This operator exports the model and schemas (input and output) in AVRO format. The exported model and schemas can be used in a TIBCO® ModelOps scoring pipeline.

Export to ModelOps operator icon

Information at a Glance

Note: This operator can only be used with TIBCO® Data Virtualization and Apache Spark 3.2 or later.

Parameter

Description
Category Tools
Data source type TIBCO® Data Virtualization
Sends output to other operators No
Data processing tool TIBCO® DV, Apache Spark 3.2 or later

Input

An input is a Modeling operator and a single tabular data set against which the model is created.

Note: The input data set must be the same as the training data set.

Algorithm

A model is created using modeling operators. Then, the model and the corresponding data set are used to generate the schema of the output data set. The model is then packaged as a zip archive that the TIBCO® ModelOps can use. The input and output schemas are converted to AVRO schema format before uploading and publishing to the configured TIBCO® ModelOps instance along with the input model.

Restrictions

The column names must be compliant with the AVRO names. A valid name can only contain ASCII characters for letters, numbers, and underscore. For more information, see the Apache Avro documentation.

Configuration

The following table provides the configuration details for the Export Model to ModelOps operator.

Parameter Description
Notes Notes or helpful information about this operator's parameter settings. When you enter content in the Notes field, a yellow asterisk appears on the operator.
TIBCO® ModelOps Project Specify the name of the TIBCO® ModelOps project where you want to export the model.
Note: The specified project must exist in TIBCO® ModelOps. If the project does not exist, then an error message is displayed.
Model File Path Specify the path of the model file for the current workflow. This creates the following artifacts in the TIBCO® ModelOps project:
  • <Model File Path>.zip

  • <Model File Path>-input.avsc

  • <Model File Path>-output.avsc

Description Enter a brief description of the exported model for the current workflow. This description appears in the model artifact in TIBCO® ModelOps.
Commit Message Enter a commit message for the model file. It provides information about the changes made to the model.

Default: Committed by TeamStudio

Overwrite? Specify whether to overwrite the existing model.
  • When set to Yes, the model is overwritten with a new model version (new published version) if the model already exists with the same name. Input and output schemas for the model are overwritten only if the schemas are different from existing ones.

  • When set to No, the model is not overwritten. If the model of the same name already exists, the operator fails with an error message that notifies the user of the existing model.

Default: Yes

Output

Visual Output
  • Summary: A text field displaying the summary of the exported model.
Output to successive operators
None. This is a terminal operator.

Example

The following example uses the golf data set to build the Naive Bayes model and then exports the model to the TIBCO® ModelOps by using the Export Model to ModelOps operator.

Export Model to ModelOps workflow
Data
golf: This data set contains the following information:
  • Multiple columns namely outlook, temperature, wind, humidity, and play.
  • Multiple rows (14 rows).
Parameter Setting
The parameter settings for the Export Model to ModelOps operator are as follows:
  • TIBCO® ModelOps Project: TeamStudioExport

  • Model File Path: integratedtest/naivebayes

  • Description: Model from Team Studio

  • Commit Message: Committed by TeamStudio

  • Overwrite?: Yes

Output
The following figure displays the result for the Export Model to ModelOps operator.
Export Model to ModelOps operator - Output

Once the model gets exported, you can see the models in the mentioned project in the TIBCO® ModelOps instance. The following figure from the TIBCO® ModelOps displays the model and its information.

Exported Models in ModelOps