test_generate_system_analysis in integration_tests.summarization_test.SummarizationTest is flaky

[tetsuok]

Sometimes, test_generate_system_analysis in integration_tests.summarization_test.SummarizationTest is failing in CI. An example failure can be found at: https://github.com/neulab/ExplainaBoard/actions/runs/3214011085/jobs/5254156113. The relevant part of the log is the following:

======================================================================
ERROR: test_generate_system_analysis (integration_tests.summarization_test.SummarizationTest)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/runner/work/ExplainaBoard/ExplainaBoard/integration_tests/summarization_test.py", line 55, in test_generate_system_analysis
    sys_info = processor.process(metadata, data, skip_failed_analyses=True)
  File "/home/runner/work/ExplainaBoard/ExplainaBoard/explainaboard/processors/processor.py", line 563, in process
    overall_statistics = self.get_overall_statistics(metadata, sys_output)
  File "/home/runner/work/ExplainaBoard/ExplainaBoard/explainaboard/processors/processor.py", line 532, in get_overall_statistics
    external_stats = self._gen_external_stats(sys_info)
  File "/home/runner/work/ExplainaBoard/ExplainaBoard/explainaboard/processors/processor.py", line 196, in _gen_external_stats
    dataset = loader.load()
  File "/home/runner/work/ExplainaBoard/ExplainaBoard/explainaboard/loaders/loader.py", line 178, in load
    dataset_loaded_data = self._dataset_file_loader.load(
  File "/home/runner/work/ExplainaBoard/ExplainaBoard/explainaboard/loaders/file_loader.py", line 410, in load
    raw_data = self.load_raw(data, source)
  File "/home/runner/work/ExplainaBoard/ExplainaBoard/explainaboard/loaders/file_loader.py", line 675, in load_raw
    dataset = load_dataset(
  File "/opt/hostedtoolcache/Python/3.9.14/x64/lib/python3.9/site-packages/datalabs/load.py", line 2144, in load_dataset
    builder_instance.download_and_prepare(
  File "/opt/hostedtoolcache/Python/3.9.14/x64/lib/python3.9/site-packages/datalabs/builder.py", line 747, in download_and_prepare
    self._download_and_prepare(
  File "/opt/hostedtoolcache/Python/3.9.14/x64/lib/python3.9/site-packages/datalabs/builder.py", line 844, in _download_and_prepare
    split_generators = self._split_generators(dl_manager, **split_generators_kwargs)
  File "/home/runner/.cache/expressai/modules/datasets_modules/datalab/cnn_dailymail/162caa8b3eb106635aa2ec0c7a3df867092817fed79936d3b3372c9296337018/cnn_dailymail.py", line 333, in _split_generators
    train_files = _subset_filenames(dl_paths, datalabs.Split.TRAIN)
  File "/home/runner/.cache/expressai/modules/datasets_modules/datalab/cnn_dailymail/162caa8b3eb106635aa2ec0c7a3df867092817fed79936d3b3372c9296337018/cnn_dailymail.py", line 210, in _subset_filenames
    cnn = _find_files(dl_paths, "cnn", urls)
  File "/home/runner/.cache/expressai/modules/datasets_modules/datalab/cnn_dailymail/162caa8b3eb106635aa2ec0c7a3df867092817fed79936d3b3372c9296337018/cnn_dailymail.py", line 189, in _find_files
    files = sorted(os.listdir(top_dir))
NotADirectoryError: [Errno 20] Not a directory: '/home/runner/.cache/expressai/datalab/downloads/1bc05d24fa6dda2468e83a73cf6dc207226e01e3c48a507ea716dc0421da583b/cnn/stories'