openml-python
openml-python copied to clipboard
openml
Python module to interface with OpenML
Currently, `list_evaluations` can only return the average of evaluations. However, the webserver also returns the evaluations per fold. It would be great to access them in python, too.
I think it should use a de-serialization function, as currently it returns all hyperparameter values as string, which is wrong IMO.
Currently neither `Pipeline` nor `ColumnTransformer` may contain two different steps with the same type of transformer. I think this should be allowed. Consider a scenario where I have a dataset...
Several users seem to experience difficulties downloading dataset `41081`. The process is killed with a sigkill signal when pickling the dataset to disk [here](https://github.com/openml/openml-python/blob/develop/openml/datasets/dataset.py#L301). We deduce this from the fact...
A continuation of #261 which was largely addressed by #748. However, certain flows and studies persist and fail to be deleted. Therefore, the file deletion can be made more strict.
When loading a documentation page on a specific header in Firefox, e.g. [usage#basic-usage](https://openml.github.io/openml-python/master/usage.html#basic-usage), when coming from a different page, the navbar hides the section header. ---- [usage#basic-usage](https://openml.github.io/openml-python/master/usage.html#basic-usage) shows as: ...
#### Description I created a virtual environment for openml, and when I tried to pip install I need to install other packages, it would be great if a requirements.txt file...
Currently our python objects tend to have many attributes that we only want modified by the API itself, e.g. all of the reference to `id`s. However, by not prefixing those...
A good unit test would consist of the following steps: 1. instantiate a classifier 2. solve a small task with a complex decision boundary 3. upload the classifier to OpenML...
Make sure that the tests ensure that functions _create_description_xml() and _create_run_from_xml() are each others exact complements. e.g., * a parameter free classifier * a classifier with multiple components * a...