rsmtool

rsmtool copied to clipboard

Now that we are Python 3.6+ only, we should switch to using format strings.

desilinguist

It's confusing that the name of the argument is `flag_column` and one of the values that it takes is also `flag_column`. This makes the docstring very confusing to write. We...

desilinguist

For classifiers that do not support expected probabilities, we currently rely on SKLL to raise a warning and proceed generating integer scores. The final report still says "Predictions analyzed in...

aloukina

Add best practices for sharing reports to documentation.

1

It would be useful to add some best practices for sharing reports with other people to the documentation. When to send just the HTML, when to zip up everything, when...

desilinguist

Output file or directory in rsmpredict

3

Currently, `rsmpredict` supports an undocumented option of specifying an output directory instead of file if the output_file does not have a `.csv` or `.xlsx` extension. However, there are several inconsistencies:...

desilinguist

aloukina

Replace all path operations and functions with `pathlib`

1

Now that we are Python 3.6+ only, it makes more sense to use the more readable `pathlib.Path` interface rather than os-level functions.

desilinguist

Look into parallelization

1

[JIRA] Perhaps, we can use multiprocessing or multithreading to speed up model training, report generation etc. This might be relevant: http://ipyparallel.readthedocs.io/en/latest/intro.html

aloukina

https://github.com/nbQA-dev/nbQA

desilinguist

It would be great to have the option to convert all NaN feature values to 0 (though fine to not have it be the default). We could show a warning...

aoifecahill