import_model_version.py fails to import a model from local enviroment

[jiropardo]

When trying to import a model from a local environment to Databricks UC using the import_model_version function, the import fails with AWS zone related issues

line 89, in _get_region_name
head_bucket_resp = temp_client.head_bucket(Bucket=self.bucket)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

MLflow version used:
mlflow-skinny==2.14.3

Function used
https://github.com/mlflow/mlflow-export-import/blob/a572e944ebee82fe9b37c5e7b35541d01c26253b/mlflow_export_import/model_version/import_model_version.py

Code for importing the model

import_model_version(
model_name=f"model_catalog.model_schema.Model_12",
experiment_name=f"/test/EXPERIMENTS/Testing_model",
input_dir=model_migration/resources/01/,
create_model=True,
import_stages_and_aliases=True,
)

The workaround to make this work is to manually specify the s3 bucket in boto3 like below

import boto3
boto3.setup_default_session(region_name='region-of-workspace')

However, this error does not occur within Databricks on a cluster with MLflow 2.14.3 or latter but when running this import function from local. The error is coming from the client code and not the server code as the import_model_version does not set the region when running it locally

Please review this and clarify whether this function needs the region to be passed explicitly as a parameter or if the current client code will be updated to set the region when run in a local environment