Small optimizations to scale and translate of Affine2D #17165
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Do you have a profiling script that shows the difference? |
import matplotlib.transforms as tr
import numpy as np
a = tr.Affine2D.from_values(1, 0, 0, 1, 5, 5)
arr = a.get_matrix().copy()
arr2 = a.get_matrix().copy()
arr3 = a.get_matrix().copy()
print(a.get_matrix())
def scale_mpl(arr, sx, sy):
scale_mtx = np.array(
[[sx, 0.0, 0.0], [0.0, sy, 0.0], [0.0, 0.0, 1.0]])
return np.dot(scale_mtx, arr, out=arr)
def scale(arr, sx, sy):
arr[0, :] *= sx
arr[1, :] *= sy
return arr
def scale_2(arr, sx, sy):
arr[0, 0] *= sx
arr[0, 1] *= sx
arr[0, 2] *= sx
arr[1, 0] *= sy
arr[1, 1] *= sy
arr[1, 2] *= sy
return arr
scale_mpl(arr, 2, 5), scale(arr2, 2, 5), scale_2(arr3, 2, 5)
# In[143]:
get_ipython().run_cell_magic('timeit', 'arr = a.get_matrix().copy()', 'scale_mpl(arr, 2, 2)\nscale_mpl(arr, 0.5, 0.5)')
# In[144]:
get_ipython().run_cell_magic('timeit', 'arr = a.get_matrix().copy()', 'scale(arr, 2, 2)\nscale(arr, 0.5, 0.5)')
# In[145]:
get_ipython().run_cell_magic('timeit', 'arr = a.get_matrix().copy()', 'scale_2(arr, 2, 2)\nscale_2(arr, 0.5, 0.5)')
# In[149]:
a = tr.Affine2D.from_values(1, 0, 0, 1, 5, 5)
arr = a.get_matrix().copy()
arr2 = a.get_matrix().copy()
arr3 = a.get_matrix().copy()
def translate_mpl(arr, tx, ty):
scale_mtx = np.array(
[[1.0, 0.0, tx], [0.0, 1.0, ty], [0.0, 0.0, 1.0]])
return np.dot(scale_mtx, arr, out=arr)
def translate(arr, tx, ty):
arr[0, 2] += tx
arr[1, 2] += ty
return arr
translate_mpl(arr, 2, 5), translate(arr2, 2, 5)
# In[150]:
get_ipython().run_cell_magic('timeit', 'arr = a.get_matrix().copy()', 'translate(arr, 2, 2)\ntranslate(arr, -2, -2)')
# In[151]:
get_ipython().run_cell_magic('timeit', 'arr = a.get_matrix().copy()', 'translate_mpl(arr, 2, 2)\ntranslate_mpl(arr, -2, -2)') |
|
CI seems to be possibly genuinely failing. |
|
Yep, it is a real issue. I think it is caused by the fact that neither the Affine2D constructor nor the get_matrix method make copy. Since transformation are initiallized by matricies of other transformation this is a problem. |
|
This issue was hidden by the fact that all transformations created a new array. |
17fc9e1 to
3f26a2f
Compare
anntzer
reviewed
Apr 17, 2020
QuLogic
reviewed
Apr 17, 2020
|
flake8 check fails with
|
timhoffm
approved these changes
Apr 18, 2020
timhoffm
previously requested changes
Apr 18, 2020
68006e7 to
b1bbb8d
Compare
timhoffm
reviewed
Apr 21, 2020
timhoffm
approved these changes
Apr 29, 2020
36d66cf to
0013958
Compare
…tion. Fix revealed bug in backend bases, save some transformation copys
0013958 to
88e40cb
Compare
anntzer
reviewed
Apr 29, 2020
7bd9d3c to
4f66d31
Compare
anntzer
reviewed
Apr 30, 2020
4f66d31 to
740ffbf
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR Summary
Instead of building a transformation matrix and do the matrix multiplication, just change the matrix directly. This change also exposed a bug in draw_path_collection where Affine2D was called with the wrong argument.
PR Checklist