Couldn't find a dataset script at /content/tsv/tsv.py or any data file in the same directory

[andysingal]

Describe the bug

---------------------------------------------------------------------------
FileNotFoundError                         Traceback (most recent call last)
[<ipython-input-48-6a7b3e847019>](https://localhost:8080/#) in <cell line: 7>()
      5 }
      6 
----> 7 csv_datasets_reloaded = load_dataset("tsv", data_files=data_files)
      8 csv_datasets_reloaded

2 frames
[/usr/local/lib/python3.10/dist-packages/datasets/load.py](https://localhost:8080/#) in dataset_module_factory(path, revision, download_config, download_mode, dynamic_modules_path, data_dir, data_files, **download_kwargs)
   1489                     raise e1 from None
   1490                 if isinstance(e1, FileNotFoundError):
-> 1491                     raise FileNotFoundError(
   1492                         f"Couldn't find a dataset script at {relative_to_absolute_path(combined_path)} or any data file in the same directory. "
   1493                         f"Couldn't find '{path}' on the Hugging Face Hub either: {type(e1).__name__}: {e1}"

FileNotFoundError: Couldn't find a dataset script at /content/tsv/tsv.py or any data file in the same directory. Couldn't find 'tsv' on the Hugging Face Hub either: FileNotFoundError: Dataset 'tsv' doesn't exist on the Hub

Steps to reproduce the bug

data_files = {
    "train": "/content/PUBHEALTH/train.tsv",
    "validation": "/content/PUBHEALTH/dev.tsv",
    "test": "/content/PUBHEALTH/test.tsv",
}

tsv_datasets_reloaded = load_dataset("tsv", data_files=data_files)
tsv_datasets_reloaded

---------------------------------------------------------------------------
FileNotFoundError                         Traceback (most recent call last)
<ipython-input-48-6a7b3e847019> in <cell line: 7>()
      5 }
      6 
----> 7 csv_datasets_reloaded = load_dataset("tsv", data_files=data_files)
      8 csv_datasets_reloaded

2 frames
/usr/local/lib/python3.10/dist-packages/datasets/load.py in dataset_module_factory(path, revision, download_config, download_mode, dynamic_modules_path, data_dir, data_files, **download_kwargs)
   1489                     raise e1 from None
   1490                 if isinstance(e1, FileNotFoundError):
-> 1491                     raise FileNotFoundError(
   1492                         f"Couldn't find a dataset script at {relative_to_absolute_path(combined_path)} or any data file in the same directory. "
   1493                         f"Couldn't find '{path}' on the Hugging Face Hub either: {type(e1).__name__}: {e1}"

FileNotFoundError: Couldn't find a dataset script at /content/tsv/tsv.py or any data file in the same directory. Couldn't find 'tsv' on the Hugging Face Hub either: FileNotFoundError: Dataset 'tsv' doesn't exist on the Hub

Expected behavior

load the data, push to hub

Environment info

jupyter notebook RTX 3090

[mariosasko]

Hi! You can load this dataset with:

data_files = {
    "train": "/content/PUBHEALTH/train.tsv",
    "validation": "/content/PUBHEALTH/dev.tsv",
    "test": "/content/PUBHEALTH/test.tsv",
}

tsv_datasets_reloaded = load_dataset("csv", data_files=data_files, sep="\t")

To support your load_dataset call, defining aliases for the packaged builders, as suggested in https://github.com/huggingface/datasets/issues/5625, must be implemented. We can consider adding this feature if more people request it.

(Also answered on the Discord here)