pose2text: reduce disk read pressure

[AmitMY]

• there is no need to check if the file exists again
• if one file has many items, it shouldn't re-read the file a million times

[GerrySant]

The version you propose may encounter errors in case a file does not exist:

Traceback (most recent call last):
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/builder.py", line 1606, in _prepare_split_single
    for key, record in generator:
  File "/Users/sant/repositories/multimodalhugs/multimodalhugs/data/datasets/pose2text.py", line 110, in _generate_examples
    dataset = dataset.map(mapping_function)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/arrow_dataset.py", line 560, in wrapper
    out: Union["Dataset", "DatasetDict"] = func(self, *args, **kwargs)
                                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/arrow_dataset.py", line 3035, in map
    for rank, done, content in Dataset._map_single(**dataset_kwargs):
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/arrow_dataset.py", line 3408, in _map_single
    example = apply_function_on_filtered_inputs(example, i, offset=offset)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/arrow_dataset.py", line 3300, in apply_function_on_filtered_inputs
    processed_inputs = function(*fn_args, *additional_args, **fn_kwargs)
                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/sant/repositories/multimodalhugs/multimodalhugs/data/datasets/pose2text.py", line 100, in mapping_function
    buffer = self._read_pose(sample['source'])
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/sant/repositories/multimodalhugs/multimodalhugs/data/datasets/pose2text.py", line 83, in _read_pose
    with open(file_path, "rb") as pose_file:
         ^^^^^^^^^^^^^^^^^^^^^
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/streaming.py", line 75, in wrapper
    return function(*args, download_config=download_config, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/utils/file_utils.py", line 947, in xopen
    return open(main_hop, mode, *args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: '/Users/sant/phd/first_paper/data/How2Sign/sentence_level/val/rgb_front/pose_estimation/eY32ru3Nstc_20-8-rgb_front.pose'

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/Users/sant/repositories/multimodalhugs/multimodalhugs/multimodalhugs_cli/training_setup.py", line 47, in <module>
    main()
  File "/Users/sant/repositories/multimodalhugs/multimodalhugs/multimodalhugs_cli/training_setup.py", line 34, in main
    pose2text_setup(args.config_path)
  File "/Users/sant/repositories/multimodalhugs/multimodalhugs/training_setup/pose2sign_training_setup.py", line 25, in main
    dataset.download_and_prepare(data_path)
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/builder.py", line 924, in download_and_prepare
    self._download_and_prepare(
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/builder.py", line 1647, in _download_and_prepare
    super()._download_and_prepare(
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/builder.py", line 999, in _download_and_prepare
    self._prepare_split(split_generator, **prepare_split_kwargs)
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/builder.py", line 1485, in _prepare_split
    for job_id, done, content in self._prepare_split_single(
  File "/Users/sant/anaconda3/envs/multimodalhugs/lib/python3.12/site-packages/datasets/builder.py", line 1642, in _prepare_split_single
    raise DatasetGenerationError("An error occurred while generating the dataset") from e
datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

So I suggest to modify this to the following code fragment:

        # Filter out samples where the file path does not exist
        dataset = dataset.filter(lambda sample: file_exists_filter('source_signal', sample))

        # Apply the update to the VIDEO_NAME column
        dataset = dataset.map(mapping_function)

[AmitMY]

Addressed. Changed the order of filter and map