Uploading datasets

[ArlindKadra]

Hey all,
I am working with dataset uploading and I stumbled upon something.

openml-python/openml/datasets/dataset.py

Lines 373 to 397 in f618d81

	def publish(self):
	"""Publish the dataset on the OpenML server.
	Upload the dataset description and dataset content to openml.
	Returns
	-------
	return_code : int
	Return code from server
	return_value : string
	xml return from server
	"""
	file_elements = {'description': self._to_xml()}
	file_dictionary = {}
	if self.data_file is not None:
	file_dictionary['dataset'] = self.data_file
	return_value = _perform_api_call("/data/", file_dictionary=file_dictionary,
	file_elements=file_elements)
	self.dataset_id = int(xmltodict.parse(return_value)['oml:upload_data_set']['oml:id'])
	return self

The function publish() in the OpenMLDataset makes use of the xml description of a dataset and an arff file to upload a dataset at OpenML. However in the way that the class is implemented right now, self.data_file is a string containing the path to the dataset file.

In my opinion we should have a method that takes the description and the arff file as an argument at openml.datasets at the functions module.

Something like:
publish_dataset(description, file)

What is your opinion regarding this?

[mfeurer]

@amueller @janvanrijn @joaquinvanschoren you might be interested in that. When looking at the suggestion, please keep in mind that neither a run nor a flow are currently created by a user but rather by a helper function.

[janvanrijn]

Short version of the answer (from phone) I agree.

Slightly longer: dataset class is a little complicated ATM and contains many things that it shouldn't have imo, e.g. reference to features, qualities, maybe even splits? Ideally this should all be decoupled (might be a bit hard, I never wanted to start this)

[mfeurer]

I agree that the class is a bit complicated, but I disagree on that it shouldn't contain the features and qualities. This makes it easy for a user to access all kinds of things related to datasets. It does not contain the splits, these are in the task.

[mfeurer]

I believe this is now implemented in #440.