You want to work with only a subset of directories from a large Git repository in Databricks. Which feature should you use?