Skip to content

[MRG] Fix ordering #139

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 4 commits into from
Apr 7, 2020
Merged

[MRG] Fix ordering #139

merged 4 commits into from
Apr 7, 2020
+51 −18

Conversation

AdrienCorenflos
Copy link
Contributor

Fixes #138

@AdrienCorenflos AdrienCorenflos mentioned this pull request Apr 1, 2020
Closed
@rflamary rflamary added the bug label Apr 2, 2020
rtavenar
rtavenar reviewed Apr 2, 2020
View reviewed changes
Copy link
Contributor

@rtavenar rtavenar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch ! I just have a minor comment. Also I am not sure if it is better to have two different tests for the variants with and without weights, but this is not very important I guess.

ot/lp/__init__.py Outdated
@@ -656,7 +656,7 @@ def emd_1d(x_a, x_b, a=None, b=None, metric='sqeuclidean', p=1., dense=True,
perm_a = np.argsort(x_a_1d)
perm_b = np.argsort(x_b_1d)

G_sorted, indices, cost = emd_1d_sorted(a, b,
G_sorted, indices, cost = emd_1d_sorted(a[perm_a.flatten()], b[perm_b.flatten()],
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I guess you do not have to flatten the permutation indices since they are computed from 1d arrays, or do you ?

Copy link
Contributor

@rtavenar rtavenar Apr 2, 2020 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could you explain why you flatten here?

Apart from that, LGTM!

rtavenar
rtavenar reviewed Apr 2, 2020
View reviewed changes
test/test_ot.py Outdated
np.testing.assert_allclose(wass, wass1d_emd2)

# check loss is similar to scipy's implementation for Euclidean metric
wass_sp = wasserstein_distance(u.reshape((-1,)), v.reshape((-1,)))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You forgot the weights here, which is probably why the test fails at the moment

@AdrienCorenflos
Copy link
Contributor Author

AdrienCorenflos commented Apr 2, 2020 via email

@rflamary
Copy link
Collaborator

rflamary commented Apr 2, 2020

Hello @AdrienCorenflos,

Thank you for finding the bug and the PR.

In addition to the comments from @rtavenar be careful to check pep8 on test_ot.py or else the tests will fail.

rtavenar
rtavenar approved these changes Apr 2, 2020
View reviewed changes
Copy link
Contributor

@rtavenar rtavenar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM ! Thanks a lot for this bugfix !

@rflamary rflamary changed the title Fix ordering [MRG] Fix ordering Apr 3, 2020
@rflamary rflamary merged commit d399f62 into PythonOT:master Apr 7, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
1 more reviewer

@rtavenar rtavenar rtavenar approved these changes

Reviewers whose approvals may not affect merge requirements
Assignees
No one assigned
Labels
bug
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Bad value in EMD 1D
3 participants
@AdrienCorenflos @rflamary @rtavenar