Passing a value to the params parameter in the batch_predict() method of TablesClient class does not have any effect.

For example, passing the value params={'feature_importance': 'true'} to batch_predict() is supposed to have feature importance generated along with the output, but there was none.

Environment details

• OS type and version: Ubuntu 18.04.5 LTS
• Python version: 3.5.2, 3.6.10, 3.7.2, 3.8.6
• pip version: 20.2.4
• google-cloud-automl version: 2.1.0

Steps to reproduce

1. Make a batch prediction with params (e.g. params={'feature_importance': 'true'}).
2. Invoke the request.
3. Notice that the generated output does not generate feature importance.

Code example

from google.cloud import automl_v1beta1 client = automl_v1beta1.TablesClient(project='my-project', region='us-central1') client.batch_predict( bigquery_input_uri='bq://my-project.dataset.table', bigquery_output_uri='bq://my-project', model_display_name='my_model', params={'feature_importance': 'true'} )