Add support for Matplotlib 2.2 #1119
Conversation
… with nans in x and y axis
|
I tried running with matplotlib 2.2 in connection with qdev-dk/qdev-wrappers#88 and ran into errors with plotting. However, turned out the problem we were trying to solve was unrelated to matplotlib version. |
|
@ThorvaldLarsen I was referring to this one which I think depend on matplotlib version and which this fixes
|
Codecov Report
@@ Coverage Diff @@
## master #1119 +/- ##
=========================================
+ Coverage 78.79% 79% +0.21%
=========================================
Files 46 46
Lines 6564 6573 +9
=========================================
+ Hits 5172 5193 +21
+ Misses 1392 1380 -12 |
qcodes/plots/qcmatplotlib.py
Outdated
| # numpy (1.14) has a deprecation warning related to shared | ||
| # masks | ||
| # lets silence this by making sure that this is not shared | ||
| # before |
There was a problem hiding this comment.
PEP8 is nice and all but IMHO this indentation is worse than too long lines 😛
There was a problem hiding this comment.
Yeah, I'm still getting used to VSCode. That would just have been meta+q in Emacs... sigh...
qcodes/plots/qcmatplotlib.py
Outdated
|
|
||
| # now shift to get edges coordinates rather than center | ||
| # coordinates | ||
| # first Add padding on both sides equal to endpoints |
There was a problem hiding this comment.
This comment is now wrong. Perhaps we should just use insert, That seems simpler than pad for this
|
|
||
| assert len(args[0]) == M + 1 | ||
| assert len(args[1]) == N + 1 | ||
|
|
There was a problem hiding this comment.
Should we not test that the content of x and y is as expected too?
There was a problem hiding this comment.
Yes... the test can/should be extended a lot. I just got very lazy.
Matplotlib no longer works with nans in the x and y coordinates for pcolormesh. This fills out any missing values by extrapolating based on the step size or the existing data, falling back to a default stepsize if there is only one line. IMHO this also makes the code for shifting coordinates from bin centers to edges slightly simpler.
@ThorvaldLarsen I think you saw this issue somewhere in the qdev wrappers.
@QCoDeS/core