Fix test_take to make axis optional when ndim == 1 #192
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
honno
approved these changes
Jul 17, 2023
There was a problem hiding this comment.
This LGTM, just have a few minor points that encroach on bikesheding 😅 I think I would write this something like
def test_take(x, data):
...
axis_strat = st.integers(0, max(x.ndim - 1, 0))
if x.ndim == 1: # i.e. axis is optional
axis_strat |= st.none()
axis = data.draw(axis_strat, label="axis")
_axis = axis or 0
_indices = data.draw(
st.lists(st.integers(0, x.shape[_axis] - 1), min_size=1, unique=True),
label="_indices",
)
indices = xp.asarray(_indices, dtype=dh.default_int)
note(f"{indices=}")
kw = data.draw(hh.specified_kwargs(("axis", axis, None)), label="kw")
out = xp.take(x, indices, **kw)
ph.assert_dtype("take", in_dtype=x.dtype, out_dtype=out.dtype)
ph.assert_shape(
"take",
out_shape=out.shape,
expected=x.shape[:_axis] + (len(_indices),) + x.shape[_axis + 1 :],
kw=dict(
x=x,
indices=indices,
axis=axis,
),
)
out_indices = sh.ndindex(out.shape)
axis_indices = list(sh.axis_ndindex(x.shape, _axis))
...- Have a core strategy that we "add"
st.none()too if relevant—if you didn't know, piping strategies is syntactic sugar forst.one_of(). - Generate
axisas the actual kwarg value, and have_axisas the processed/explicit logical value. - We can utilise
hh.specified_kwargs()for when we want to test not passing the kwarg.test_linspaceis a good example of its use.
I say all this because most of the existing tests are written to follow the above coding patterns—we don't have to be consistent or anything, just good to be aware of the prior work.
|
I mentioned it in the OP, but I intentionally didn't test axis=None, because it's not clear to me that we actually want that keyword default, given that it's the same as axis=0 for the current semantics. Libraries like NumPy use |
I didn't explicitly test axis=None because it's not clear to me that should actually be supported, given that that's the same as axis=0.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I didn't explicitly test axis=None because it's not clear to me that should actually be supported, given that that's the same as axis=0.
Fixes #191