Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Multivariate plotting in imshow, pcolor and pcolormesh#29221

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to ourterms of service andprivacy statement. We’ll occasionally send you account related emails.

Already on GitHub?Sign in to your account

Open
trygvrad wants to merge12 commits intomatplotlib:main
base:main
Choose a base branch
Loading
fromtrygvrad:multivariate-norm

Conversation

trygvrad
Copy link
Contributor

PR summary

This PR continues the work of#28658 and#28454, aiming toclose#14168.

Featuresincluded in this PR:

  1. AMultiNorm class. This is a subclass ofcolors.Normalize and holdsn_variate norms.
  2. Exposes the functionality provided byMultiNorm together withBivarColormap andMultivarColormap to the plotting functionsaxes.imshow(...),axes.pcolor, and `axes.pcolormesh(...)
import matplotlib.pyplot as pltimport numpy as npim_A = np.arange(500)[np.newaxis, :]*np.ones((100, 500))im_B = np.arange(100)[:, np.newaxis]*np.ones((100, 500))im_A[:, :] = np.sin(im_A**0.5)**2im_B[:, :] = np.sin(im_B**0.5)**2fig, axes = plt.subplots(3, 1, figsize=(8, 6))cm = axes[0].imshow(im_A, cmap='grey', vmin=0, vmax=1)axes[0].set_title('A')cm = axes[1].imshow(im_B, cmap='grey', vmin=0, vmax=1)axes[1].set_title('B')cm = axes[2].imshow((im_A, im_B),                    cmap='BiOrangeBlue',                    vmin=(0, 0), vmax=(1, 1))axes[2].set_title('A & B')for ax in axes:    ax.set_xticks([])    ax.set_yticks([])

image

Featuresnot included in this PR:

  • Multivariate equivalent offig.colorbar(...). I will make a draft for this, but I may need help with regards to the layout engine.
  • Examples. Some documentation is ready on a separate branch, seethis, andthis, but the PR that includes these will be made after we find afig.colorbar(...) solution for bivariate/multivariate colormaps.
  • A selection of colormaps. I have written about thishere andhere. You can see the proposed colormaps/documentationhere and [here](file:///home/trygvrad/trygvrad.github.io/mpl_docs/Bivariate colormap reference.html), but we will need a separate discussion on these (in particular regarding the names)
  • The multivariate implementation foraxes.imshow(...) included here only acceptsinterpolation_stage=='rgba', and notinterpolation_stage=='data'.

PR checklist

@trygvrad
Copy link
ContributorAuthor

Also: this does not (yet) include changes to the docstrings, i.e. the following will need to change:

"""Parameters----------X : array-like or PIL image    The image data. Supported array shapes are:    - (M, N): an image with scalar data. The values are mapped to      colors using normalization and a colormap. See parameters *norm*,      *cmap*, *vmin*, *vmax*.    - (M, N, 3): an image with RGB values (0-1 float or 0-255 int).    - (M, N, 4): an image with RGBA values (0-1 float or 0-255 int),      i.e. including transparency.    The first two dimensions (M, N) define the rows and columns of    the image.    Out-of-range RGB(A) values are clipped."""

to include a line like

"""etc.    - (K, M, N): multiple images with scalar data when combined with a    multivariate or bivariate colormap using a MultiNorm."""

@trygvrad
Copy link
ContributorAuthor

I updated the docstrings, so now this PR should be ready for review.

There is an error with the docs on circleci, but I haven't found a link between the error and the changes made in this PR, so I believe it to be unrelated.

@tacaswelltacaswell modified the milestones:v3.10.0-doc,v3.11.0Jan 23, 2025
@trygvrad
Copy link
ContributorAuthor

trygvrad commentedFeb 8, 2025•
edited
Loading

@story645
Thank you for the feedback :)
I believe I have addressed all the comments now (I'll mark them as resolved in the coming days).

also: this will need a squash-merge

# pass scalar data
# and already formatted data
return data
elif data.dtype in [np.complex64, np.complex128]:
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

Option 1 please.

if A.ndim == 3 and A.shape[-1] == 1:
A = A.squeeze(-1) # If just (M, N, 1), assume scalar and apply colormap.
if not (A.ndim == 2 or A.ndim == 3 and A.shape[-1] in [3, 4]):
if A.ndim == 3 and A.shape[0] == 2:
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

to verify, these are here as a hint to the user to use multi-variate color maps and data that would have failed before would still fail now?

Copy link
ContributorAuthor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

Yes, that is the intention

raise TypeError(f"Image data of dtype {A.dtype} cannot be "
f"converted to float")
else:
for key in A.dtype.fields:
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

I think this fails to catchn_input >1 + non-float-like + non-structured data?

Copy link
ContributorAuthor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

This might be true.
I'm trying to think of a data type that can not be cast to float and is not structured in order to test this

If you try something with data type object, it depends a bit on the shape, all the cases I have tried have been caught:

fig,ax=plt.subplots()im=np.empty((5,5),dtype='object')ax.imshow(im,cmap='BiOrangeBlue')>ValueError:Invaliddataentryformutlivariatedata.Thedatamustcontaincomplexnumbers,orhaveafirstdimension2,orbeofadtypewith2fieldsim=np.empty((2,5,5),dtype='object')ax.imshow(im,cmap='BiOrangeBlue')>TypeError:Imagedataofdtype [('f0','O'), ('f1','O')]cannotbeconvertedtoasequenceoffloats

Where the first error originates incolorizer._ensure_multivariate_data()

Do you have some idea of other tests I should run, or do we want to have specific tests for_normalize_image_array()?

"""
Check validity of image-like input *A* and normalize it to a format suitable for
Image subclasses.
"""
A = mcolorizer._ensure_multivariate_data(n_input, A)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

minor thing. bit jarring that sometimes the API inf(n_input, A) and othertimesf(A, n_input). I think we should preferf(A, n_input) everywhere for internal consistency and no-API change on existing functions (like the one we are in now).

Copy link
ContributorAuthor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

This is a good comment :) I will try to address it soon

Copy link
ContributorAuthor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

I changed the call signature ofcolorizer._ensure_multivariate_data(...) tocolorizer._ensure_multivariate_data(A, n_input)

I looked to see if there were any other functions with a similar signature which should be flipped, but I did not find any in my quick search.

@story645
Copy link
Member

If you've got the bandwidth,@greglucas can you give this a look (skim) since you've been working on colormapping?

@trygvrad
Copy link
ContributorAuthor

@greglucas It would be great to know if/when you will have time to look at this :)
Just let us know if you don't have the bandwidth right now, and we can poke someone else

@trygvrad
Copy link
ContributorAuthor

trygvrad commentedMar 27, 2025•
edited
Loading

@story645@tacaswell would you two be willing to finish your review of this?
Or perhaps@ksunden?

EDIT: my apologies@story645 I didn't see you had just assigned this when I wrote my comment :)

@story645
Copy link
Member

@anntzer asked if the more cleanup/refactor/not directly necessary for implementation parts of this PR could be spun off into a different set and if this PR could get rebased to clean up the commit history.

@anntzer
Copy link
Contributor

(Just to give an example, I'm currently splitting off#29807 into a bunch of smaller preparatory PRs:#29817,#29828,#29838,#29843,#29829.)

@trygvrad
Copy link
ContributorAuthor

trygvrad commentedApr 1, 2025•
edited
Loading

@anntzer I'll see what I can do.
I thinking of splitting it as:

1. MultiNorm class2a. Add support for multivariate pcolor and pcolormesh + tests 2b. Add support for multivariate imshow + tests

I'll also see if there are some misc changes that can be a separate PR between 1 and 2.
In any case, 1. will have to be merged before 2a and b. Would you like me to make all the branches at once, or does it make more sense to make the PR for 2a and b once 1. is merged?

@anntzer
Copy link
Contributor

I guess it doesn't really matter, because we know the end result thanks to this PR anyways.
(I won't be able to review anything before next week, in any case.)

trygvrad reacted with thumbs up emoji

@trygvrad
Copy link
ContributorAuthor

@anntzer@story645 I opened PRs:
#29876 → The MultiNorm class
#29877 → Other changes needed to expose the functionality.

Hopefully this is easier to review :)

anntzer reacted with thumbs up emojistory645 reacted with heart emoji

trygvradand others added12 commitsApril 13, 2025 13:13
If an image is shown, and the user hovers over the image, the value at that coordinate should be displayed. The number formating is handled by colorizer.Colorizer._format_cursor_data_override(). This is now updated to also handle the case where a MultiNorm is used (i.e. a bivariat or multivariate colormap), in which case multiple values are displayed.
The previous version of this test passed for numpy 1.x but not for 2.x. This should pass for both
Co-authored-by: Thomas A Caswell <tcaswell@gmail.com>
changed from  colorizer._ensure_multivariate_data(n_input, A)to  colorizer._ensure_multivariate_data(A, n_input)
Sign up for freeto join this conversation on GitHub. Already have an account?Sign in to comment
Reviewers

@tacaswelltacaswelltacaswell left review comments

@story645story645story645 left review comments

At least 1 approving review is required to merge this pull request.

Assignees

@story645story645

@anntzeranntzer

Projects
None yet
Milestone
v3.11.0
Development

Successfully merging this pull request may close these issues.

Feature request: Bivariate colormapping
4 participants
@trygvrad@story645@anntzer@tacaswell
[8]ページ先頭
©2009-2025 Movatter.jp