Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules. #4115

trexfeathers · 2021-05-04T14:50:11Z

🚀 Pull Request

Description

Restores output of the anomaly log colouring gallery to pre #4087 by using a new linscale argument (can't go back to linscale=0 as no longer allowed by Matplotlib).

Consult Iris pull request check list

rcomer · 2021-05-04T14:59:36Z

docs/gallery_code/general/plot_anomaly_log_colouring.py

        vmin=-maximum_scale_level,
        vmax=maximum_scale_level,
    )
    # Setting "linthresh=minimum_log_level" makes its non-logarithmic
    # data range equal to our 'zero band'.
-    # Setting "linscale=1" maps the whole zero band to the middle colour value
+    # Setting "linscale=0.01" maps the whole zero band to the middle colour value
    # (i.e., 0.5), which is the neutral point of a "diverging" style colormap.


I'll be honest, I'm really struggling to match this comment up with what the mpl docs say about linscale. I'm getting the feeling that either the original comment was wrong, or the behaviour has changed. Or I'm just missing something... 😕

It's the first bit about "linthresh=minimum_log_level" that I'm particularly struggling with, rather than the changed bit.

The language of these comment lines makes no sense to me either, but I assumed that to be due to the writer's greater knowledge in this area. Seems important to understand and correct before merging.

Looks like @pp-mo was the originator, but a Whole Seven Years Ago!!! #1040

Given how much matplotlib would have moved on in seven years, it's entirely possible that there are better ways of achieving this outcome now anyway. So possibly we should just merge this now for the sake of fixing master, and open a new issue to revisit later?

Having read in more detail and talked it over with @stephenworsley, I'm reasonably happy with how it's written already.

Just found this.
FWIW there was a perfectly good documentation for this up to Matplotlib v3.3,
https://matplotlib.org/3.3.4/api/_as_gen/matplotlib.colors.SymLogNorm.html

But they seem to have lost it somewhere since ??

update : It's not unique : a lot of the classes have lost their docstring details.
I'm trying to work out how to raise this as an issue.

trexfeathers · 2021-05-04T15:22:34Z

The good news is that this finally gives us a passing py3.8 doc tests (gallery).

jamesp

Looks good to me, the only failing tests now are the square root failures 👍

…s. (SciTools#4115)

* mpl 3.4.1 updates (#4087) * replace most recent hashes (#4112) * Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules. (#4115) * Cartopy 0.19 updates (#4128) * Use assertArrayAllClose for sqrt test (#4118) * using AllClose for sqrt test * Omitting the checksum from test cml * use ArrayAllClose (rebase reset it?) * Iris py38 (#3976) * support for py38 * update CI and noxfile * enforce alphabetical xml element attribute order * full tests for py38 + fix docs-tests * add whatsnew entry * update doc-strings + review actions * Alternate xml handling routine (#29) * all xml tests pass for nox tests-3.8 * restored docstrings * move sort_xml_attrs * make sort_xml_attrs a classmethod * update sort_xml_attr doc-string Co-authored-by: Bill Little <[email protected]> * add jamesp to whatsnew + minor tweak Co-authored-by: James Penn <[email protected]> * reinstate black and nox links * Linkcheck update (#4104) * Updated links * Added login remark * Removed extra space * change to kick cirrus * kick cirrus * test verbose on cirrus * Removed test settings. Co-authored-by: Bill Little <[email protected]> Co-authored-by: Martin Yeo <[email protected]> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: James Penn <[email protected]> Co-authored-by: James Penn <[email protected]> Co-authored-by: tkknight <[email protected]>

* master: (49 commits) Update CI environment lockfiles + Cartopy 0.19 changes (SciTools#4125) separate arg types from descriptions (SciTools#4100) Use assertArrayAllClose for sqrt test (SciTools#4118) Removed branch suffix (SciTools#4117) Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules. (SciTools#4115) replace most recent hashes (SciTools#4112) Linkcheck update (SciTools#4104) CI lockfiles (SciTools#4108) Fix bug in gallery_tests runner (SciTools#4111) add logo to conda and pypi badges (SciTools#4088) pre-commit-ci update (SciTools#4092) Fix cirrus ci and mpl (SciTools#4087) Update readme docs stable (SciTools#4089) linux kernel version fix for conda 4.10+ (SciTools#4084) update docs pypi instructions (SciTools#4077) update flake8 pre-commit (SciTools#4067) Add GitHub discussions badge (SciTools#4070) conda requirements add pip (SciTools#4062) add pre-commit.ci badge and doc support (SciTools#4061) cirrus-ci credits for non-master (SciTools#4060) ...

* master: (87 commits) Contourf Antialias Fix (SciTools#4150) minor tidy of cirrus and nox (SciTools#4152) update bug and feature templates (SciTools#4147) Update CI environment lockfiles + Cartopy 0.19 changes (SciTools#4125) separate arg types from descriptions (SciTools#4100) Use assertArrayAllClose for sqrt test (SciTools#4118) Removed branch suffix (SciTools#4117) Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules. (SciTools#4115) replace most recent hashes (SciTools#4112) Linkcheck update (SciTools#4104) CI lockfiles (SciTools#4108) Fix bug in gallery_tests runner (SciTools#4111) add logo to conda and pypi badges (SciTools#4088) pre-commit-ci update (SciTools#4092) Fix cirrus ci and mpl (SciTools#4087) Update readme docs stable (SciTools#4089) linux kernel version fix for conda 4.10+ (SciTools#4084) update docs pypi instructions (SciTools#4077) update flake8 pre-commit (SciTools#4067) Add GitHub discussions badge (SciTools#4070) ...

…t-iris-imagehash#44.

…s: Duplicates SciTools#4115.

@trexfeathers

* Add release highlights and pin rc version (#3898) * Add release highlights and pin rc version * review actions * reorder release highlights (#3899) Tweak release highlights * Add whatsnew announcement (#3900) * Fix spelling (#3903) * Fix unit label handling (#3902) * Add failing test of plotting * Implement fix to pass test * Update idiff to ignore irrelevant hyphens in path * Update imagerepo (following docs) * Update after review by @trexfeathers * Add whatsnew entries * Move whatsnew entries into correct file * Release Docs Improvements (#3895) * Minor phrasing change in 'Release candidate'. * Before release deprecations. * Whatsnew highlights section. * Relax setup.py setup requirements (#3909) * Updated CF saver version in User Guide and docstring (#3925) * Updated CF saver version in User Guide and docstring * Remove references to CF version of the loader in docstrings * Added whatsnew * Pin cftime<1.3.0 * Migrate to cirrus-ci (#3928) * migrate from travis-ci to cirrus-ci * added whatsnew entries * ignore url for doc link check (#3929) * whatsnew for coord default units (#3924) * Cube._summary_coord_extra: efficiency and bugfix (#3922) * Add Documentation Title Case Capitalization (#3940) * Use Title Case Capitalisation for Documentation * add whatsnew enter * CI requirements drop pip packages (#3939) * requirements pip to conda * use pip install over develop * default PY_VER to python versions * update links (#3942) * update links * added s to http * Add support for 1-d weights in collapse. (#3943) * Remove warning for convert_units on lazy data (#3951) * drop stickler references in docs (#3953) * drop stickler references in docs * remove sticker from common links * update docs for travis-ci to cirrus-ci (#3954) * update docs for travis-ci to cirrus-ci * add 'travis-ci' reference locally to whatsnew * update whatsnew comment * docs for nox (#3955) * docs for nox * add titles, notices and additional detail * review actions * Resolve test coverage (#3947) * test coverage for __init__ and __call__ * test coverage for metadata resolve and coverage * partial test coverage for metadata mapping * python 3.6 workaround for deepcopy of mock.sentinel * test coverage for Resolve._free_mapping * test coverage for Resolve convenience methods * add test stub for Resolve._metadata_mapping * fix Test__tgt_cube_position * test coverage for shape * test coverage for _as_compatible_cubes * test coverage for Resolve._metadata_mapping * test coverage for Resolve._prepare_common_dim_payload * test coverage for Resolve._prepare_common_aux_payload * test coverage for Resolve._prepare_points_and_bounds * test coverage for Resolve._create_prepared_item * test coverage for Resolve._prepare_local_payload_dim * test coverage for Resolve._prepare_local_payload_aux * test coverage for Resolve._prepare_local_payload_scalar + docs URL skip * test coverage for Resolve._prepare_local_payload * test coverage for Resolve._metadata_prepare * added docs URL linkcheck skip * test coverage for Resolve._prepare_factory_payload * test coverage for Resolve._get_prepared_item * review actions * test coverage for Resolve.cube * pin v3.0.0 version and whatnew date (#3956) * update github ci checks image (#3957) * Promote unknown units to dimensionless in aux factories (#3965) * promote unknown to dimensionless units in aux factories * patch aux factories to promote unknown to dimensionless units for formula terms * add whatnew PR for entry * Release branch prepare for v3.0.2 (#4044) * update intersphinx mapping and matplotlib urls (#4003) * update intersphinx mapping and matplotlib urls * use matplotlib intersphinx where possible * review actions * review actions * cirrus-ci compute credits (#4007) * cirrus-ci conditional tasks (#4019) * cirrus-ci conditional tasks * use bc for bash arithmetic * revert back to sed * use expr * reword * minor documentation changes * review actions * prepare v3.0.2 release * Fix test_incompatible_dimensions test (#3977) * test_incompatible_dimensions used a ragged array for the test, which has been deprecated in numpy, and now fails if dtype is anything other than object. This test appears to be checking that the addition of a [2x4] masked array to a [2x3] masked cube should raise a ValueError. This commit fixes the creation of `data3` object to be a [2x4] non-ragged array. * Added entry to what's new * Added name to core developer list :) * Update latest.rst Fixed space in PR macro call * update whatsnew v3.0.2 Co-authored-by: James Penn <[email protected]> * um_stash_source attribute improved handling (#4035) * Modified pyke rule * Tests added * Black and whatsnew * Include PR number * Remove latest.rst * Add what's new * Support for py38 and Cartopy 0.19 (#4130) * mpl 3.4.1 updates (#4087) * replace most recent hashes (#4112) * Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules. (#4115) * Cartopy 0.19 updates (#4128) * Use assertArrayAllClose for sqrt test (#4118) * using AllClose for sqrt test * Omitting the checksum from test cml * use ArrayAllClose (rebase reset it?) * Iris py38 (#3976) * support for py38 * update CI and noxfile * enforce alphabetical xml element attribute order * full tests for py38 + fix docs-tests * add whatsnew entry * update doc-strings + review actions * Alternate xml handling routine (#29) * all xml tests pass for nox tests-3.8 * restored docstrings * move sort_xml_attrs * make sort_xml_attrs a classmethod * update sort_xml_attr doc-string Co-authored-by: Bill Little <[email protected]> * add jamesp to whatsnew + minor tweak Co-authored-by: James Penn <[email protected]> * reinstate black and nox links * Linkcheck update (#4104) * Updated links * Added login remark * Removed extra space * change to kick cirrus * kick cirrus * test verbose on cirrus * Removed test settings. Co-authored-by: Bill Little <[email protected]> Co-authored-by: Martin Yeo <[email protected]> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: James Penn <[email protected]> Co-authored-by: James Penn <[email protected]> Co-authored-by: tkknight <[email protected]> * Intersection bounds fix (replacement PR) (#4059) * fix intersection out of bounds point * fix intersection out of bounds point take 2 * add referencing comment * bootstrap ci for 3.0.x (#4154) * add cirrus-ci docker support * wip * wip * wip * wip * wip * add pip * revert docker * fix add_weekday (#32) * fix add_weekday * fix black link * fix link Co-authored-by: Ruth Comer <[email protected]> * Tweak to speed up dask wrapping of netcdf variables (#4135) * Use 'meta' in da.from_array to stop it sampling netcdf variables, which is quite slow. * Fix PR number. * Fix test. * Review changes. * Update docs/iris/src/whatsnew/3.0.2.rst Co-authored-by: lbdreyer <[email protected]> * Fb fix cube coord arithmetic (#4159) * fix coord with cube arithmetic * add coord arithmetic test coverage * add a whatsnew entry * Pp daskfix (#4141) * Remove workaround when dask-wrapping PP data, obsoleted by #4135. * Remove old slice testing * Add whats new * move whitespace? * missing line Co-authored-by: lbdreyer <[email protected]> * add release date to v3.0.2 whatsnew (#4160) * update readme logo img src and href (#4006) (#4216) * In cube.intersection, find split cells using a tolerant equality check (#4220) * In cube.intersection, find split cells using a tolerant equality check * remove unused var * Add bounds check to tests; add what's new * Fix whats new; update version number * Update 3.0.3.rst Update release date in `3.0.3.rst` * Wide cubestr fix 3v0vx v2 (#4233) * Widen cube printout for long ancil or cell-measure names. * Adjust result for fixed cube-units printout. * Added whatsnew. * Include newest whatsnew in index. * Review: fix whatsnew structure. * Fix initial sections display. * Fix some typos in the cube maths docs (#4248) * (More of) Wide cubestr fix 3v0vx v2 (#4238) * Fix whatsnew for #4233. * Move 'empty slicings' whatsnew entry to 3.0.2 section. * unpin cftime (#4222) * Test wrangling vs v3.0.x (#4249) * test changes * add awol init * rename system test cases * fix PartialDateTime tests * fix typo * add whatsnew * Fix mergeback PR #4035 * Fix mergeback PR #4035 tests * Update mergeback nox conda-lock files Co-authored-by: tkknight <[email protected]> Co-authored-by: Zeb Nicholls <[email protected]> Co-authored-by: Martin Yeo <[email protected]> Co-authored-by: Jon Seddon <[email protected]> Co-authored-by: Ruth Comer <[email protected]> Co-authored-by: Patrick Peglar <[email protected]> Co-authored-by: James Penn <[email protected]> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: James Penn <[email protected]> Co-authored-by: Ruth Comer <[email protected]> Co-authored-by: lbdreyer <[email protected]> Co-authored-by: lbdreyer <[email protected]>

trexfeathers added the Experience: Medium label May 4, 2021

trexfeathers requested review from jamesp and rcomer May 4, 2021 14:50

Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules.

ad09e2a

trexfeathers force-pushed the fix_anomaly_linscale branch from 561e568 to ad09e2a Compare May 4, 2021 14:53

rcomer reviewed May 4, 2021

View reviewed changes

jamesp approved these changes May 4, 2021

View reviewed changes

jamesp merged commit 055a60b into SciTools:master May 4, 2021

trexfeathers mentioned this pull request May 4, 2021

Undo changes from #40 (anomaly log plot) as these were not the appropriate fix. SciTools/test-iris-imagehash#44

Merged

rcomer mentioned this pull request May 5, 2021

Use assertArrayAllClose for sqrt test #4118

Merged

rcomer pushed a commit to rcomer/iris that referenced this pull request May 12, 2021

Corrected plot_anomaly_log_colouring for new Matplotlib linscale rule…

707bda7

…s. (SciTools#4115)

pp-mo added a commit to pp-mo/iris that referenced this pull request Jul 6, 2021

Remove ref from imagerepo.json, matching SciTools#4115, to please tes…

e2ac80c

…t-iris-imagehash#44.

trexfeathers mentioned this pull request Jul 6, 2021

Default connectivity start_index #4224

Closed

pp-mo pushed a commit to pp-mo/iris that referenced this pull request Jul 7, 2021

Corrected plot_anomaly_log_colouring for new Matplotlib linscale rule…

8dfec01

…s: Duplicates SciTools#4115.

trexfeathers deleted the fix_anomaly_linscale branch August 31, 2021 15:44

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules. #4115

Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules. #4115

trexfeathers commented May 4, 2021

rcomer May 4, 2021 •

edited

Loading

rcomer May 4, 2021

trexfeathers May 4, 2021

rcomer May 4, 2021

rcomer May 4, 2021

trexfeathers May 4, 2021

pp-mo Jul 6, 2021 •

edited

Loading

trexfeathers commented May 4, 2021

jamesp left a comment

Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules. #4115

Corrected plot_anomaly_log_colouring for new Matplotlib linscale rules. #4115

Conversation

trexfeathers commented May 4, 2021

🚀 Pull Request

Description

rcomer May 4, 2021 • edited Loading

Choose a reason for hiding this comment

rcomer May 4, 2021

Choose a reason for hiding this comment

trexfeathers May 4, 2021

Choose a reason for hiding this comment

rcomer May 4, 2021

Choose a reason for hiding this comment

rcomer May 4, 2021

Choose a reason for hiding this comment

trexfeathers May 4, 2021

Choose a reason for hiding this comment

pp-mo Jul 6, 2021 • edited Loading

Choose a reason for hiding this comment

trexfeathers commented May 4, 2021

jamesp left a comment

Choose a reason for hiding this comment

rcomer May 4, 2021 •

edited

Loading

pp-mo Jul 6, 2021 •

edited

Loading