[BUG] load_UCR_UEA_dataset checks for existence of files rather than just directories

[TonyBagnall]

What does this implement/fix? Explain your changes.

This PR does three things to code that is designed to work with the specific data format and file name structure used in time series classification. This will not impact generic data loading, it is for the specific use case of using load_UCR_UEA_dataset to load .ts files.

1. Changes the example usage for load_UCR_UEA_dataset so that it loads a baked in dataset (ArrowHead) rather than download one from timeseriesclassification.com. This is to remove the danger of problems with timeseriesclassification impacting the testing of examples
2. Private function _list_available_datasets now checks that the correct <problem_name>.ts files exist, rather than just check that the <problem_name> directory exists. This fixes a bug where an empty directory <problem_name> would cause load_UCR_UEA_dataset to think the data was stored locally, and thus crash.
3. Adds a unit test for load_UCR_UEA_dataset (although only tests on baked in data sets)

Does your contribution introduce a new dependency? If yes, which one?

No

What should a reviewer concentrate their feedback on?

load_UCR_UEA_dataset

PR checklist

For all contributions

• I've added myself to the list of contributors.
• Optionally, I've updated sktime's CODEOWNERS to receive notifications about future changes to these files.
• I've added unit tests and made sure they pass locally.
• [ x] The PR title starts with either [ENH], [MNT], [DOC], or [BUG] indicating whether the PR topic is related to enhancement, maintenance, documentation, or bug.

For new estimators

• I've added the estimator to the online documentation.
• I've updated the existing example notebooks or provided a new one to showcase how my estimator works.

[lmmentel]

Thanks, I started seeing the related error also in other PRs.