Thanks to visit codestin.com
Credit goes to github.com

Skip to content

PGF Backend: Support interpolation='none' #6792

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.

Sign up for GitHub

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

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Aug 27, 2016
Merged

PGF Backend: Support interpolation='none' #6792

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
468a977
PGF backend: skip drawing of empty images
f0k Jul 18, 2016
4bcb38f
PGF backend: support affine image transformations
f0k Jul 18, 2016
d931283
Add more transforms to affine image transform demo
f0k Jul 18, 2016
0eb17a4
RendererBase: improve draw_image() docstring
f0k Jul 20, 2016
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
RendererBase: improve draw_image() docstring
  • Loading branch information
@f0k
f0k committed Aug 26, 2016
commit 0eb17a4c4b709172c6488400412cd982f4ae2cc9
34 changes: 19 additions & 15 deletions lib/matplotlib/backend_bases.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -508,30 +508,34 @@ def get_image_magnification(self):
"""
return 1.0

def draw_image(self, gc, x, y, im, trans=None):
def draw_image(self, gc, x, y, im, transform=None):
Copy link
Member

Choose a reason for hiding this comment

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

Please do not change kwarg name, it is a (subtle) api break for anyone who is using **dd to pass in kwargs.

Copy link
Contributor Author

@f0k f0k Jul 26, 2016 •
edited
Loading

Choose a reason for hiding this comment

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

The point is that if anybody passes trans= as a kwarg, it is already broken:

$ grep -re 'def draw_image' lib/matplotlib
lib/matplotlib/backends/backend_pgf.py:    def draw_image(self, gc, x, y, im, transform=None):
lib/matplotlib/backends/backend_template.py:    def draw_image(self, gc, x, y, im):
lib/matplotlib/backends/backend_ps.py:    def draw_image(self, gc, x, y, im, transform=None):
lib/matplotlib/backends/backend_svg.py:    def draw_image(self, gc, x, y, im, transform=None):
lib/matplotlib/backends/backend_gdk.py:    def draw_image(self, gc, x, y, im):
lib/matplotlib/backends/backend_pdf.py:    def draw_image(self, gc, x, y, im, transform=None):
lib/matplotlib/backends/backend_wx.py:    def draw_image(self, gc, x, y, im):
lib/matplotlib/backends/backend_cairo.py:    def draw_image(self, gc, x, y, im):
lib/matplotlib/backend_bases.py:    def draw_image(self, gc, x, y, im, trans=None):

The base class is the only one calling it trans, which leaves a trap for anybody using **dd to pass in the keyword arguments.

Copy link
Member

Choose a reason for hiding this comment

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

That is fun.

attn @mdboom I am inclined to accept this change and an API change entry for it.

Copy link
Member

Choose a reason for hiding this comment

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

Agreed. Let's fix this to be consistent (transform everywhere) and add a note in api_changes.rst.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

and add a note in api_changes.rst.

If I see correctly, it's already covered by https://github.com/matplotlib/matplotlib/blob/master/doc/api/api_changes/2015-12-30_draw_image.rst -- if I added another note, there would be two API changes for draw_image() listed for matplotlib 2.0. So shall we just leave it at that?

Copy link
Member

Choose a reason for hiding this comment

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

eh, yeah there is an entry for it, but just amend it to note this particular issue, as it isn't obvious from the current item that there is this change as well.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Well, from the current item it isn't obvious at all what has changed:

The draw_image method implemented by backends has changed its interface.

This change is only relevant if the backend declares that it is able to transform images by returning True from option_scale_image. See the draw_image docstring for more information.

Both sentences apply to my PR as well, which just changes the changed interface. Sorry, I'm not trying to be a nuisance, but I cannot come up with a way that explicitly mentions my change without sounding silly (and confusing).

Also note that the interface change in that note belongs to #5718, which added the trans keyword in the first place, in an attempt to document the interface already present in the subclasses:
https://github.com/matplotlib/matplotlib/pull/5718/files#diff-504c288d675954e4e24de83a19242f4dR518

I'm just changing this to correctly call it transform, which has been used in subclasses even before #5718 was proposed: https://github.com/matplotlib/matplotlib/pull/5718/files#diff-2623790aa8493ba82ef737fe0e63274dL1601

So this is not actually an API change. Nobody can have used trans= as a keyword argument, because it wouldn't have worked.

Is this convincing?

Copy link
Member

Choose a reason for hiding this comment

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

Ah, I see your point about trans. So, you are merely correcting a mistake before it reaches a release. Point taken.

"""
Draw the image instance into the current axes;
Draw an RGBA image.
Copy link
Contributor Author

@f0k f0k Jul 20, 2016 •
edited
Loading

Choose a reason for hiding this comment

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

Explanation: It's not an "image instance", and the renderer does not have any notion of axes (as far as I can see). draw_image() is used both for axes and figure images.


*gc*
a GraphicsContext containing clipping information
a :class:`GraphicsContextBase` instance with clipping information.

*x*
is the distance in pixels from the left hand side of the canvas.
the distance in physical units (i.e., dots or pixels) from the left
hand side of the canvas.

*y*
the distance from the origin. That is, if origin is
upper, y is the distance from top. If origin is lower, y
is the distance from bottom
the distance in physical units (i.e., dots or pixels) from the
bottom side of the canvas.

*im*
An NxMx4 array of RGBA pixels (of dtype uint8).

*trans*
If the concrete backend is written such that
`option_scale_image` returns `True`, an affine
transformation may also be passed to `draw_image`. The
backend should apply the transformation to the image
before applying the translation of `x` and `y`.
*transform*
If and only if the concrete backend is written such that
:meth:`option_scale_image` returns ``True``, an affine
transformation *may* be passed to :meth:`draw_image`. It takes the
form of a :class:`~matplotlib.transforms.Affine2DBase` instance.
The translation vector of the transformation is given in physical
units (i.e., dots or pixels). Note that the transformation does not
override `x` and `y`, and has to be applied *before* translating
the result by `x` and `y` (this can be accomplished by adding `x`
and `y` to the translation vector defined by `transform`).
"""
raise NotImplementedError

Expand All @@ -544,8 +548,8 @@ def option_image_nocomposite(self):

def option_scale_image(self):
"""
override this method for renderers that support arbitrary
scaling of image (most of the vector backend).
override this method for renderers that support arbitrary affine
transformations in :meth:`draw_image` (most vector backends).
"""
return False

Expand Down