[MRG+1] BUG: MultiLabelBinarizer.fit_transform sometimes returns an invalid CSR matrix #7750
Conversation
This file contains hidden or 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
See scipy/scipy#6719 for context. The gist is that the `inverse` array may have a different dtype than `yt.indices`, which causes trouble down the line because, in those cases, `yt.indices` and `yt.indptr` have different dtypes. Alternately, we could insert `yt.check_format(full_check=False)` after modifying the sparse matrix members.
|
Thanks. can you add a test please? |
Older versions don't support kwargs for `astype`
jnothman
reviewed
Oct 25, 2016
| @@ -732,7 +732,7 @@ def fit_transform(self, y): | |||
| class_mapping = np.empty(len(tmp), dtype=dtype) | |||
| class_mapping[:] = tmp | |||
| self.classes_, inverse = np.unique(class_mapping, return_inverse=True) | |||
| yt.indices = inverse[yt.indices].astype(yt.indices.dtype, copy=False) | |||
There was a problem hiding this comment.
you could use np.asarray(..., dtype=) to more closely reflect this operation, I think
|
Yes, it looks like that's the sort of test we should perform wherever we do CSR manipulation. Hacky hacky hack hack. Thanks @perimosocordiae. |
|
Tests pass now. Should I squash the commits? |
|
no need. |
|
Maybe add a comment why the assert is needed and why the line is needed? It's a bit non-obvious to me. |
|
Opened #7762 to track the overall problem. |
[ci skip]
[ci skip]
|
Okay, comments added. I skipped the full CI treatment for the comment-only changes. Conversion to LIL format would test the symptom but not the cause of the error. It's possible that we may add checks in scipy to deal with this in the future, so I'd prefer to not rely on |
|
thanks :) |
amueller
pushed a commit
to amueller/scikit-learn
that referenced
this pull request
Nov 9, 2016
…nvalid CSR matrix (scikit-learn#7750) * BUG: MultiLabelBinarizer makes invalid CSR matrix See scipy/scipy#6719 for context. The gist is that the `inverse` array may have a different dtype than `yt.indices`, which causes trouble down the line because, in those cases, `yt.indices` and `yt.indptr` have different dtypes. Alternately, we could insert `yt.check_format(full_check=False)` after modifying the sparse matrix members. * Fixing for old numpy Older versions don't support kwargs for `astype` * Adding tests * line-wrapping * adding comment to tests [ci skip] * added rationale comment [ci skip]
sergeyf
pushed a commit
to sergeyf/scikit-learn
that referenced
this pull request
Feb 28, 2017
…nvalid CSR matrix (scikit-learn#7750) * BUG: MultiLabelBinarizer makes invalid CSR matrix See scipy/scipy#6719 for context. The gist is that the `inverse` array may have a different dtype than `yt.indices`, which causes trouble down the line because, in those cases, `yt.indices` and `yt.indptr` have different dtypes. Alternately, we could insert `yt.check_format(full_check=False)` after modifying the sparse matrix members. * Fixing for old numpy Older versions don't support kwargs for `astype` * Adding tests * line-wrapping * adding comment to tests [ci skip] * added rationale comment [ci skip]
Sundrique
pushed a commit
to Sundrique/scikit-learn
that referenced
this pull request
Jun 14, 2017
…nvalid CSR matrix (scikit-learn#7750) * BUG: MultiLabelBinarizer makes invalid CSR matrix See scipy/scipy#6719 for context. The gist is that the `inverse` array may have a different dtype than `yt.indices`, which causes trouble down the line because, in those cases, `yt.indices` and `yt.indptr` have different dtypes. Alternately, we could insert `yt.check_format(full_check=False)` after modifying the sparse matrix members. * Fixing for old numpy Older versions don't support kwargs for `astype` * Adding tests * line-wrapping * adding comment to tests [ci skip] * added rationale comment [ci skip]
paulha
pushed a commit
to paulha/scikit-learn
that referenced
this pull request
Aug 19, 2017
…nvalid CSR matrix (scikit-learn#7750) * BUG: MultiLabelBinarizer makes invalid CSR matrix See scipy/scipy#6719 for context. The gist is that the `inverse` array may have a different dtype than `yt.indices`, which causes trouble down the line because, in those cases, `yt.indices` and `yt.indptr` have different dtypes. Alternately, we could insert `yt.check_format(full_check=False)` after modifying the sparse matrix members. * Fixing for old numpy Older versions don't support kwargs for `astype` * Adding tests * line-wrapping * adding comment to tests [ci skip] * added rationale comment [ci skip]
maskani-moh
pushed a commit
to maskani-moh/scikit-learn
that referenced
this pull request
Nov 15, 2017
…nvalid CSR matrix (scikit-learn#7750) * BUG: MultiLabelBinarizer makes invalid CSR matrix See scipy/scipy#6719 for context. The gist is that the `inverse` array may have a different dtype than `yt.indices`, which causes trouble down the line because, in those cases, `yt.indices` and `yt.indptr` have different dtypes. Alternately, we could insert `yt.check_format(full_check=False)` after modifying the sparse matrix members. * Fixing for old numpy Older versions don't support kwargs for `astype` * Adding tests * line-wrapping * adding comment to tests [ci skip] * added rationale comment [ci skip]
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.
See scipy/scipy#6719 for context.
The gist is that the
inversearray may have a different dtype thanyt.indices, which causes trouble down the line because, in those cases,yt.indicesandyt.indptrhave different dtypes.Alternately, we could insert
yt.check_format(full_check=False)after modifying the sparse matrix members.