MyPy cleanup and absolute imports in pandas.core.dtypes.common #21008
Conversation
WillAyd
changed the title
| ABCIndexClass, ABCDateOffset) | ||
| from .inference import is_string_like, is_list_like | ||
| from .inference import * # noqa | ||
| from pandas.core.dtypes.dtypes import ( |
There was a problem hiding this comment.
There are quite some other places where we use relative imports as well (but of course not consistently ..), but so why change those and not others?
There was a problem hiding this comment.
It wasn’t the relative import that mypy was complaining about as much as the starred import. I just figured I’d convert the relative to absolute while I was touching the module
jorisvandenbossche
changed the title
Codecov Report
@@ Coverage Diff @@
## master #21008 +/- ##
==========================================
- Coverage 91.9% 91.9% -0.01%
==========================================
Files 153 153
Lines 49549 49548 -1
==========================================
- Hits 45539 45538 -1
Misses 4010 4010
Continue to review full report at Codecov.
|
jreback
added
Dtype Conversions
|
looks fine. can you rebase just to be sure. merge on green. |
Starting the conversation towards #14468
There were a couple misplaced annotations below docstrings that were causing mypy to choke when running:
A starred import in pandas.core.dtypes.common was responsible for a lot of complaints from mypy. These were being intentionally suppressed by flake8 but I figure it makes sense to clean up and convert into absolute imports