Update image tutorial. by anntzer · Pull Request #23889 · matplotlib/matplotlib

anntzer · 2022-09-14T16:39:25Z

Use Pillow to load the image rather than the discouraged plt.imread.
Remove discussion about rescaling to 0-1 as dropping the use of plt.imread also gets rid of that (not so nice) behavior. Also drop the lengthy discussion about uint8 and float32, which seems out of place here (if anything it should be in the Pillow docs). Make corresponding fixes to code to use 0-255 instead of 0-1.
Fix subplot management to go full pyplot (the local switch to OO-style is a bit weird).
Fix interpolation discussion as the default is now "nearest" (actually it's now "antialiased", but let's sweep that under the rug for this tutorial for now).

See #22790. Probably the tutorial could be further improved, but let's consider this a small step in that direction.

PR Summary

PR Checklist

Tests and Styling

Has pytest style unit tests (and pytest passes).
Is Flake 8 compliant (install flake8-docstrings and run flake8 --docstring-convention=all).

Documentation

New features are documented, with examples if plot related.
New features have an entry in doc/users/next_whats_new/ (follow instructions in README.rst there).
API changes documented in doc/api/next_api_changes/ (follow instructions in README.rst there).
Documentation is sphinx and numpydoc compliant (the docs should build without error).

story645 · 2022-09-14T16:48:17Z

+# Here, we used Pillow to open an image (with `PIL.Image.open`), and
+# immediately converted the `PIL.Image.Image` object into a 8-bit
+# (``dtype=uint8``) numpy array.


Suggested change

# Here, we used Pillow to open an image (with `PIL.Image.open`), and

# immediately converted the `PIL.Image.Image` object into a 8-bit

# (``dtype=uint8``) numpy array.

# We use Pillow to open an image (with `PIL.Image.open`), and

# immediately convert the `PIL.Image.Image` object into an 8-bit

# (``dtype=uint8``) numpy array.

And this should maybe be before the code instead of and here we go

story645 · 2022-09-14T16:50:07Z

@@ -205,20 +199,19 @@
 #
 # You can specify the clim in the call to ``plot``.


Suggested change

# You can specify the clim in the call to ``plot``.

# You can specify the clim (colormap limits) in the call to ``plot``.

I think clim is colormap limits?

story645 · 2022-09-14T16:52:03Z

-ax = fig.add_subplot(1, 2, 1)
+plt.subplot(1, 2, 1)
 imgplot = plt.imshow(lum_img)
-ax.set_title('Before')
-plt.colorbar(ticks=[0.1, 0.3, 0.5, 0.7], orientation='horizontal')
-ax = fig.add_subplot(1, 2, 2)
+plt.title('Before')
+plt.colorbar(orientation='horizontal')
+plt.subplot(1, 2, 2)
 imgplot = plt.imshow(lum_img)
-imgplot.set_clim(0.0, 0.7)
-ax.set_title('After')
-plt.colorbar(ticks=[0.1, 0.3, 0.5, 0.7], orientation='horizontal')
+imgplot.set_clim(0, 175)
+plt.title('After')
+plt.colorbar(orientation='horizontal')


I think this should stay OO since that's what we usually recommend, but with an fig, (ax1m ax2) = plt.subplots( ncols=2)

all the other examples are pyplot 'cause it's only a line of code and I think that's a fairly clear distinction?

I just got rid of the subplots, keeping only the second plot seems good enough.

story645 · 2022-09-14T16:52:38Z


 ###############################################################################
-# Here we have the default interpolation, bilinear, since we did not
+# Here we have the default interpolation ("nearest"), since we did not


Suggested change

# Here we have the default interpolation ("nearest"), since we did not

# Here we use the default interpolation ("nearest"), since we did not

timhoffm · 2022-09-14T21:02:48Z

@@ -205,20 +199,19 @@
 #
 # You can specify the clim in the call to ``plot``.


And its imshownotplot`, we are talking about.

Then maybe:

Suggested change

# You can specify the clim in the call to ``plot``.

# You can specify the colormap limits using ``imshow(..., clim=...)``.

Sorry, to comment on code you're not responsible for, but this is too glaringly wrong.

oscargus · 2022-09-15T06:25:38Z

-from PIL import Image
-
 img = Image.open('../../doc/_static/stinkbug.png')
 img.thumbnail((64, 64), Image.ANTIALIAS)  # resizes image in-place


Suggested change

img.thumbnail((64, 64), Image.ANTIALIAS) # resizes image in-place

img.thumbnail((64, 64)) # resizes image in-place

May just as well put that here instead of #23881 (which can be closed then).

I merged yours.

anntzer · 2022-09-15T09:01:01Z

Thanks, handled all comments (got rid of the subplots, as noted re: @story645's suggestion).

- Use Pillow to load the image rather than the discouraged plt.imread. - Remove discussion about rescaling to 0-1 as dropping the use of plt.imread also gets rid of that (not so nice) behavior. Also drop the lengthy discussion about uint8 and float32, which seems out of place here (if anything it should be in the Pillow docs). Make corresponding fixes to code to use 0-255 instead of 0-1. - Fix subplot management to go full pyplot (the local switch to OO-style is a bit weird). - Fix interpolation discussion as the default is now "nearest" (actually it's now "antialiased", but let's sweep that under the rug for this tutorial for now).

anntzer added Documentation topic: images labels Sep 14, 2022

story645 reviewed Sep 14, 2022

View reviewed changes

tacaswell added this to the v3.7.0 milestone Sep 14, 2022

timhoffm reviewed Sep 14, 2022

View reviewed changes

oscargus reviewed Sep 15, 2022

View reviewed changes

anntzer force-pushed the imtut branch from 2e9fe2d to cb71f65 Compare September 15, 2022 09:00

timhoffm approved these changes Sep 15, 2022

View reviewed changes

timhoffm merged commit 2ff4dc9 into matplotlib:main Sep 15, 2022

anntzer deleted the imtut branch September 15, 2022 12:36

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Update image tutorial.#23889

Update image tutorial.#23889
timhoffm merged 1 commit intomatplotlib:mainfrom
anntzer:imtut

anntzer commented Sep 14, 2022

Uh oh!

story645 Sep 14, 2022

Uh oh!

story645 Sep 14, 2022

Uh oh!

story645 Sep 14, 2022

Uh oh!

anntzer Sep 15, 2022

Uh oh!

story645 Sep 14, 2022

Uh oh!

timhoffm Sep 14, 2022

Uh oh!

oscargus Sep 15, 2022

Uh oh!

anntzer Sep 15, 2022

Uh oh!

anntzer commented Sep 15, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

		@@ -205,20 +199,19 @@
		#
		# You can specify the clim in the call to ``plot``.

	# You can specify the clim in the call to ``plot``.
	# You can specify the clim (colormap limits) in the call to ``plot``.

	# Here we have the default interpolation ("nearest"), since we did not
	# Here we use the default interpolation ("nearest"), since we did not

	# You can specify the clim in the call to ``plot``.
	# You can specify the colormap limits using ``imshow(..., clim=...)``.

	img.thumbnail((64, 64), Image.ANTIALIAS) # resizes image in-place
	img.thumbnail((64, 64)) # resizes image in-place

Uh oh!

Conversation

anntzer commented Sep 14, 2022

PR Summary

PR Checklist

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

anntzer commented Sep 15, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants