Fix matplotlib converter registering warning #26770

TomAugspurger · 2019-06-10T17:11:09Z

Tests were being skipped, because the module name changed in https://github.com/pandas-dev/pandas/pull/18307/files.

Closes pandas-dev#26760 Tests were being skipped, because the module name changed.

codecov · 2019-06-10T17:46:00Z

Codecov Report

Merging #26770 into master will increase coverage by 0.09%.
The diff coverage is 85.71%.

@@ Coverage Diff @@ ## master #26770 +/- ## ========================================== + Coverage 91.71% 91.81% +0.09%  ========================================== Files 178 178 Lines 50755 50762 +7 ========================================== + Hits 46552 46606 +54  + Misses 4203 4156 -47

Flag	Coverage Δ
#multiple	`90.4% <85.71%> (+0.1%)`	⬆️
#single	`41.91% <28.57%> (+0.6%)`	⬆️

Impacted Files	Coverage Δ
pandas/plotting/_matplotlib/core.py	`88.13% <100%> (+0.01%)`	⬆️
pandas/plotting/_matplotlib/boxplot.py	`75.46% <100%> (+0.34%)`	⬆️
pandas/plotting/_matplotlib/hist.py	`76.82% <66.66%> (-0.19%)`	⬇️
pandas/io/gbq.py	`78.94% <0%> (-10.53%)`	⬇️
pandas/core/frame.py	`96.88% <0%> (-0.12%)`	⬇️
pandas/core/config_init.py	`96.85% <0%> (+3.93%)`	⬆️
pandas/plotting/_misc.py	`64.55% <0%> (+5.06%)`	⬆️
pandas/plotting/_matplotlib/converter.py	`63.66% <0%> (+5.23%)`	⬆️
pandas/tseries/converter.py	`100% <0%> (+100%)`	⬆️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 959e799...fb694da. Read the comment docs.

codecov · 2019-06-10T17:46:06Z

Codecov Report

Merging #26770 into master will increase coverage by 0.08%.
The diff coverage is 71.87%.

@@ Coverage Diff @@ ## master #26770 +/- ## ========================================== + Coverage 91.87% 91.95% +0.08%  ========================================== Files 180 180 Lines 50712 50727 +15 ========================================== + Hits 46590 46645 +55  + Misses 4122 4082 -40

Flag	Coverage Δ
#multiple	`90.54% <71.87%> (+0.08%)`	⬆️
#single	`41.81% <25%> (+0.63%)`	⬆️

Impacted Files	Coverage Δ
pandas/plotting/_matplotlib/boxplot.py	`76.05% <100%> (+0.45%)`	⬆️
pandas/plotting/_matplotlib/tools.py	`78.53% <100%> (+0.12%)`	⬆️
pandas/plotting/_matplotlib/timeseries.py	`66.32% <100%> (ø)`	⬆️
pandas/plotting/_matplotlib/style.py	`85.96% <100%> (ø)`	⬆️
pandas/plotting/_matplotlib/__init__.py	`100% <100%> (ø)`	⬆️
pandas/util/_test_decorators.py	`93.22% <100%> (ø)`	⬆️
pandas/plotting/_matplotlib/misc.py	`38.2% <16.66%> (-0.73%)`	⬇️
pandas/plotting/_core.py	`88.2% <50%> (-0.75%)`	⬇️
pandas/plotting/_matplotlib/hist.py	`76.82% <75%> (-0.19%)`	⬇️
pandas/plotting/_matplotlib/core.py	`88.02% <75%> (-0.09%)`	⬇️
... and 6 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update baa77c3...28aaecf. Read the comment docs.

jreback · 2019-06-10T21:52:07Z

is there any way to actually test this? can run a subprocess tests and check for warnings in the output?

jorisvandenbossche · 2019-06-11T06:33:22Z

is there any way to actually test this?

@TomAugspurger actually proposed a way to test this in #24964, but you were against adding it (the test was called separately from ci/run_tests.sh. Although that indeed adds complexity to the CI testing code, this is only temporary and in a place (run_tests.sh) that does not need to be touched often. So personally, I would just add it like in the original version of that PR.

jreback · 2019-06-11T12:55:58Z

is there any way to actually test this?

@TomAugspurger actually proposed a way to test this in #24964, but you were against adding it (the test was called separately from ci/run_tests.sh. Although that indeed adds complexity to the CI testing code, this is only temporary and in a place (run_tests.sh) that does not need to be touched often. So personally, I would just add it like in the original version of that PR.

re-looking that's a good idea actually. @TomAugspurger can you re-add something like that?

TomAugspurger · 2019-06-11T13:04:06Z

Gonna try subprocess first, and then that if subprocess fails.

…

On Tue, Jun 11, 2019 at 7:56 AM Jeff Reback ***@***.***> wrote: is there any way to actually test this? @TomAugspurger <https://github.com/TomAugspurger> actually proposed a way to test this in #24964 <#24964>, but you were against adding it (the test was called separately from ci/run_tests.sh. Although that indeed adds complexity to the CI testing code, this is only temporary and in a place (run_tests.sh) that does not need to be touched often. So personally, I would just add it like in the original version of that PR. re-looking that's a good idea actually. @TomAugspurger <https://github.com/TomAugspurger> can you re-add something like that? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#26770?email_source=notifications&email_token=AAKAOITMUOOXFGNFBI6PNTTPZ6OGRA5CNFSM4HWVLID2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXNASZA#issuecomment-500828516>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAKAOIV4HWUIALZKG5BXULDPZ6OGRANCNFSM4HWVLIDQ> .

TomAugspurger · 2019-06-11T13:33:46Z

Subprocess seems to work pretty well here.

jorisvandenbossche · 2019-06-11T19:11:08Z

cc @datapythonista

jorisvandenbossche · 2019-06-13T10:07:45Z

The failures seem related (as they are about plotting), but not sure what change in this PR caused those

TomAugspurger · 2019-06-17T14:50:31Z

For posterity, my current hypothesis is that

import pandas imports matplotlib.pyplot by default again
This likely happens before we every call td.skip_if_no_mpl()
Which means we've imported matplotlib.pyplot before calling matplotlib.use(), which is bad.

d03a41d may have a warning in the log.

Not sure how to fix yet. Possibly, import pandas would only import matplotlib (not matplotlib.pyplot). Not sure if that's feasible.

jorisvandenbossche · 2019-06-17T15:07:00Z

import pandas imports matplotlib.pyplot by default again

But this is something we can fix? I suppose this was introduced in the PR with the matplotlib split

TomAugspurger · 2019-06-17T15:15:24Z

Yep. 28aaecf got them all I think.

TomAugspurger · 2019-06-17T15:16:10Z

The warning was being emitted by matplotlib: https://travis-ci.org/pandas-dev/pandas/jobs/546786281#L2428 (that's before 28aaecf)

TomAugspurger · 2019-06-17T16:26:48Z

OK, all green if you have another chance to look @jorisvandenbossche.

Basically, we can / need to have import pandas do import matplotlib. But it can't do import matplotlib.pyplot. Those need to be delayed. I've added that to our code checks.

And again, this is all going away in a release or so.

In my debugging, I've compat code like https://github.com/pandas-dev/pandas/pull/26770/files#diff-53e8fd9a1f847dd40d9aba697909409fL524, which no longer seem to be tested, and https://github.com/pandas-dev/pandas/pull/26770/files#diff-53e8fd9a1f847dd40d9aba697909409fL1096, which isn't needed for any recent matplotlib version.

jreback · 2019-06-21T02:06:57Z

thanks @TomAugspurger

merged, so @datapythonista can rebase the backend PR

Fixed converter warning.

fb694da

Closes pandas-dev#26760 Tests were being skipped, because the module name changed.

jreback added this to the 0.25.0 milestone Jun 10, 2019

jreback added the Visualization plotting label Jun 10, 2019

jorisvandenbossche changed the title ~~Fixed converter warning.~~ Fix matplotlib converter registering warning Jun 11, 2019

TomAugspurger added 2 commits June 11, 2019 08:04

Merge remote-tracking branch 'upstream/master' into 26760-converter

9ff10fe

subprocess test

15800e3

cleanup import route

1af0ba5

jorisvandenbossche mentioned this pull request Jun 12, 2019

DOC: Fixing some more warnings #26810

Merged

4 tasks

TomAugspurger added 2 commits June 13, 2019 10:52

draw

c12dc7c

Merge remote-tracking branch 'upstream/master' into 26760-converter

6f71319

jorisvandenbossche mentioned this pull request Jun 17, 2019

DOC: remove temporary okwarning/okexcept #26843

Closed

3 tasks

TomAugspurger added 2 commits June 17, 2019 08:43

Merge remote-tracking branch 'upstream/master' into 26760-converter

96043d3

surface warning

d03a41d

delay pyplot import

28aaecf

jreback merged commit 4850b28 into pandas-dev:master Jun 21, 2019

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Fix matplotlib converter registering warning #26770

Fix matplotlib converter registering warning #26770

Uh oh!

TomAugspurger commented Jun 10, 2019 •

edited

Loading

codecov bot commented Jun 10, 2019

codecov bot commented Jun 10, 2019 •

edited

Loading

jreback commented Jun 10, 2019

jorisvandenbossche commented Jun 11, 2019

jreback commented Jun 11, 2019

TomAugspurger commented Jun 11, 2019 via email

TomAugspurger commented Jun 11, 2019

jorisvandenbossche commented Jun 11, 2019

jorisvandenbossche commented Jun 13, 2019

TomAugspurger commented Jun 17, 2019 •

edited

Loading

jorisvandenbossche commented Jun 17, 2019

TomAugspurger commented Jun 17, 2019

TomAugspurger commented Jun 17, 2019

TomAugspurger commented Jun 17, 2019

jreback commented Jun 21, 2019

Labels

3 participants

Uh oh!

Fix matplotlib converter registering warning #26770

Fix matplotlib converter registering warning #26770

Uh oh!

Conversation

TomAugspurger commented Jun 10, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

codecov bot commented Jun 10, 2019

Codecov Report

codecov bot commented Jun 10, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

jreback commented Jun 10, 2019

jorisvandenbossche commented Jun 11, 2019

jreback commented Jun 11, 2019

TomAugspurger commented Jun 11, 2019 via email

TomAugspurger commented Jun 11, 2019

jorisvandenbossche commented Jun 11, 2019

jorisvandenbossche commented Jun 13, 2019

TomAugspurger commented Jun 17, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

jorisvandenbossche commented Jun 17, 2019

TomAugspurger commented Jun 17, 2019

TomAugspurger commented Jun 17, 2019

TomAugspurger commented Jun 17, 2019

jreback commented Jun 21, 2019

Labels

3 participants

TomAugspurger commented Jun 10, 2019 •

edited

Loading

codecov bot commented Jun 10, 2019 •

edited

Loading

TomAugspurger commented Jun 17, 2019 •

edited

Loading