You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
check_array has several parameters that just enable a check on a property of the array, like ensure_2d, ensure_min_samples, ... They have no effect on the output array: they just have the effect to raise an error or not. They usually have the naming pattern ensure_xxx which I think is intuitive and explicit.
force_all_finite is another example of such behavior but doesn't follow the same naming pattern. I think it should be renamed ensure_all_finite.
it would make the current set of params more consistent, intuitive and self explanatory.
it would allow to add new params with the naming pattern force_xxx, that have a different behavior e.g. have an effect on the output array, without bringing confusion. This is for instance the case in FEA Add writeable parameter to check_array #29018 that proposes to add force_writeable.
check_array
has several parameters that just enable a check on a property of the array, likeensure_2d
,ensure_min_samples
, ... They have no effect on the output array: they just have the effect to raise an error or not. They usually have the naming patternensure_xxx
which I think is intuitive and explicit.force_all_finite
is another example of such behavior but doesn't follow the same naming pattern. I think it should be renamedensure_all_finite
.force_xxx
, that have a different behavior e.g. have an effect on the output array, without bringing confusion. This is for instance the case in FEA Add writeable parameter to check_array #29018 that proposes to addforce_writeable
.cc @thomasjpfan
The text was updated successfully, but these errors were encountered: