Roman Yurchak
Roman Yurchak
Follow up on https://github.com/giotto-ai/giotto-tda/pull/367 this time from an upstream branch to check as secrets are not shared to forks. Just checking that git clone with an SSH key works for...
Opening this issue to consider using [sphinx-gallery](https://github.com/sphinx-gallery/sphinx-gallery) for examples as opposed to jupyter notebook. The basic format is .py files interlaced with .rst blocks for description, but once rendered corresponding...
Follow up on https://github.com/giotto-ai/giotto-tda/issues/109. Currently nightly builds are named `giotto-tda-nightly` and are uploaded [on PyPi](https://pypi.org/project/giotto-learn-nightly/). This issue aims to discuss whether the approach is optimal, with respect to, - using...
test_projection_values_equal_slice failed on master in fcffaf8 due to unreliable test timings in hypothesis, ``` =================================== FAILURES =================================== ______________________ test_projection_values_equal_slice ______________________ > ??? self = message = 'Hypothesis test_projection_values_equal_slice(X=array([[0.]])) produces unreliable...
Generally libraries linked against numpy are forward compatible, meaning that wheels should be generated with minimal numpy versions (here. `1.17`) and they will be compatible with following numpy versions. The...
Both [pybind11](https://pypi.org/project/pybind11/) and [cmake](https://pypi.org/project/cmake/) are on PyPi and conda. Can we not specify them as regular dependencies instead of manually installing in setup.py which leads to issues like https://github.com/giotto-ai/giotto-tda/issues/308 ?...
I think the current situation where both giotto-tda, giotto-lean, giotto-learn-nightly and giotto-tda-nightly can be confusing. For instance, I just searched for "giotto nightly" and ended up installing the wrong version...
As discussed earlier it might be worth investigating the use of [scikit-build](https://github.com/scikit-build/scikit-build) for integration with Cmake, instead of the custom logic currently defined in setup.py. I have not used it...
One way to make CI faster could be cache installation or compilation objects with ccache in CI. Cf https://docs.microsoft.com/en-us/azure/devops/pipelines/caching/?view=azure-devops for more details. I will look into it. Another way mentioned...
If we what to be more in sync with CI used in scikit-learn, scikit-learn-extra, imbalanced-learn etc.