ENH:made default tick formatter to switch to scientific notation earlier #14832

sameshl · 2019-07-17T14:56:34Z

The default tick formatter switches to scientific notation only for very small values,
but there is a range of slightly bigger values where scientific notation would be better too.

changed the formatter limits from -7,7 to -5,5

closes #14800

PR Summary

PR Checklist

Has Pytest style unit tests
Code is Flake 8 compliant
New features are documented, with examples if plot related
Documentation is sphinx and numpydoc compliant
Added an entry to doc/users/next_whats_new/ if major new feature (follow instructions in README.rst there)
Documented in doc/api/api_changes.rst if API changed in a backward-incompatible way

The default tick formatter switches to scientific notation only for very small values, but there is a range of slightly bigger values where scientific notation would be better too. changed the formatter limits from -7,7 to -5,5 closes #14800

jklymak · 2019-07-17T16:17:53Z

This prob needs an API change note, and you probably need to edit the default .matplotlibrc?

Add note in api section and changed new defaults in matplotlibrc

jklymak · 2019-07-17T18:02:50Z

doc/api/prev_api_changes/api_changes_3.1.0.rst

@@ -1157,3 +1157,8 @@ result in undefined behavior.  It now throws a `ValueError`.

 The signature of the (private) ``Axis._update_ticks`` has been changed to not
 take the renderer as argument anymore (that argument is unused).
+
+Reduced default value of `axes.formatter.limits`


Sorry, should have been clear about where this goes. it gets a new file in doc/api/next-api-changes/

@jklymak I am not able to exactly understand the naming convention of the files in doc/api/next-api-changes/. For eg, in 2019-05-08-AL.rst what exactly does AL mean?

AL are the authors initials. Which is a bit mysterious, i.e. in that case "Antony Lee".

And the date is the date of modification? Right?

Yes; rough is fine.

Moved api change note from prev_api_change/api_changes/3.1.0.rst to next_api_changes/2019-07-18-SL.rst

timhoffm · 2019-07-18T19:50:11Z

There are two failing image comparison tests because of the change. How do we want to handle these?

I propose to set the defaults back to -7, 7 for these tests. It's not worth or necessary regenerating the images.

sameshl · 2019-07-19T14:09:01Z

I propose to set the defaults back to -7, 7 for these tests

@timhoffm Do you mean revert the change in matplotlibrc to -5, 5 ?

timhoffm · 2019-07-21T08:51:55Z

Yes, at the beginning of these two tests, just add

# use former defaults to match existing baseline image
plt.rcParams['axes.formatter.limits'] = -7, 7

Note: There's no need to revert that setting at the end of the test because each test starts with the default settings.

sameshl · 2019-07-21T09:09:34Z

I see. I made the required changes. Thanks for the info @timhoffm!

timhoffm · 2019-07-21T18:07:39Z

Sorry, I may not have been clear enough.

I don't think we should apply this to all image comparison tests, but only on a case-by-case basis in the actual test functions where it's necessary to get the the tests pass without regenerating the baseline images (IIRC that have been two tests).

sameshl · 2019-07-22T14:14:00Z

Sorry, I think I misunderstood you. But thank you for helping me out. 👍 I think I have made the required changes

QuLogic

The code looks fine, but I don't know about the defaults change.

tacaswell · 2019-07-25T21:18:05Z

Seems reasonable, but has anyone thought about what this does to large numbers? We might want to have asymmetric thresholds.

timhoffm · 2019-07-27T06:35:49Z

@tacaswell Good point. Maybe -5, 6? 5000000 is not quite readable anymore either.

sameshl · 2019-08-01T14:24:24Z

@timhoffm I think -5, 6 would be appropriate. I can make that change

timhoffm

Blocking to prevent merge before the large-number limit is resolved.

We have a lot of approvals for the current state of -5, 5. But I would prefer -5, 6. Do we agree on that?

jklymak · 2019-08-06T16:24:19Z

... piling on: It'd be nice for a test w the new limits. It would also clarify the new behaviour and why this limit is better, because otherwise I'm not sure I follow.

timhoffm

My issues have been addressed. We need a second re-approval.

@jklymak What exactly would you like to test. I wouldn‘t want to create an image for this.

timhoffm · 2019-08-11T10:04:04Z

ping @jklymak do you need additional tests?

tacaswell · 2019-08-12T20:18:45Z

doc/api/next_api_changes/2019-07-18-SL.rst

+Reduced default value of `axes.formatter.limits`
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Changed the default value of `axes.formatter.limits` from -7, 7 to -5, 6 for better readability.


@sameshl If you get a chance could you add a .. plot example to the API change note? It greatly helps people who are skimming the release note to understand the change!

tacaswell

Modulo an example plot in the API change note.

sameshl · 2019-08-13T05:14:47Z

@tacaswell I will update api note with the .. plot example by tomorrow

tacaswell · 2019-08-22T23:21:57Z

I will open a PR with the example, merged to keep things moving.

Thanks @sameshl !

sameshl · 2019-08-26T04:40:12Z

Thanks for taking up the example @tacaswell !

tacaswell · 2019-08-27T17:43:28Z

Done in #15137

anntzer added the API: default changes label Jul 17, 2019

jklymak added this to the v3.2.0 milestone Jul 17, 2019

ENH:made default tick formatter to switch to scientific notation earlier

75437b2

Add note in api section and changed new defaults in matplotlibrc

jklymak reviewed Jul 17, 2019

View reviewed changes

ENH:made default tick formatter to switch to scientific notation earlier

8df22a7

Moved api change note from prev_api_change/api_changes/3.1.0.rst to next_api_changes/2019-07-18-SL.rst

Changed defaults back to -5, 5 for failing test cases

925da2b

timhoffm approved these changes Jul 22, 2019

View reviewed changes

QuLogic approved these changes Jul 25, 2019

View reviewed changes

QuLogic requested a review from tacaswell July 25, 2019 08:36

dstansby approved these changes Aug 6, 2019

View reviewed changes

timhoffm requested changes Aug 6, 2019

View reviewed changes

changed the formatter limits from -5,5 to -5,6

289284a

timhoffm approved these changes Aug 7, 2019

View reviewed changes

jklymak self-requested a review August 12, 2019 20:18

tacaswell reviewed Aug 12, 2019

View reviewed changes

tacaswell approved these changes Aug 12, 2019

View reviewed changes

tacaswell merged commit ac0a423 into matplotlib:master Aug 22, 2019

sameshl deleted the tick_formatter_limits branch August 23, 2019 04:34

Uh oh!

ENH:made default tick formatter to switch to scientific notation earlier #14832

ENH:made default tick formatter to switch to scientific notation earlier #14832

Uh oh!

Conversation

sameshl commented Jul 17, 2019

PR Summary

PR Checklist

Uh oh!

jklymak commented Jul 17, 2019

Uh oh!

jklymak Jul 17, 2019

Choose a reason for hiding this comment

Uh oh!

sameshl Jul 18, 2019

Choose a reason for hiding this comment

Uh oh!

jklymak Jul 18, 2019

Choose a reason for hiding this comment

Uh oh!

sameshl Jul 18, 2019

Choose a reason for hiding this comment

Uh oh!

jklymak Jul 18, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

timhoffm commented Jul 18, 2019

Uh oh!

sameshl commented Jul 19, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

timhoffm commented Jul 21, 2019

Uh oh!

sameshl commented Jul 21, 2019

Uh oh!

timhoffm commented Jul 21, 2019

Uh oh!

sameshl commented Jul 22, 2019

Uh oh!

QuLogic left a comment

Choose a reason for hiding this comment

Uh oh!

tacaswell commented Jul 25, 2019

Uh oh!

timhoffm commented Jul 27, 2019

Uh oh!

sameshl commented Aug 1, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

timhoffm left a comment

Choose a reason for hiding this comment

Uh oh!

jklymak commented Aug 6, 2019

Uh oh!

timhoffm left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

timhoffm commented Aug 11, 2019

Uh oh!

tacaswell Aug 12, 2019

Choose a reason for hiding this comment

Uh oh!

tacaswell left a comment

Choose a reason for hiding this comment

Uh oh!

sameshl commented Aug 13, 2019

Uh oh!

tacaswell commented Aug 22, 2019

Uh oh!

sameshl commented Aug 26, 2019

Uh oh!

tacaswell commented Aug 27, 2019

Uh oh!

Uh oh!

jklymak Jul 18, 2019 •

edited

Loading

sameshl commented Jul 19, 2019 •

edited

Loading

sameshl commented Aug 1, 2019 •

edited

Loading

timhoffm left a comment •

edited

Loading