Load test images from scipy.datasets

The API `scipy.misc.accent()` is deprecated. Use the new API
`scipy.datasets.accent()`.

README.md

@@ -2,8 +2,8 @@ ProxImaL
 =====================
 ![Build Status](https://github.com/comp-imaging/ProxImaL/actions/workflows/python-package.yml/badge.svg)
 
-ProxImaL is a Python-embedded modeling language for image optimization problems. 
-It allows you to express your problem in a natural way that follows the math, 
+ProxImaL is a Python-embedded modeling language for image optimization problems.
+It allows you to express your problem in a natural way that follows the math,
 and automatically determines an efficient method for solving the problem.
 ProxImaL makes it easy to experiment with many different priors and other problem reformulations,
 without worrying about the details of how the problem is solved.
@@ -17,7 +17,7 @@ import scipy.misc
 import matplotlib.pyplot as plt
 
 # Generate data.
-I = scipy.misc.ascent()
+I = scipy.datasets.ascent()
 np.random.seed(1)
 b = I + 10*np.random.randn(*I.shape)
 
@@ -53,5 +53,5 @@ See [the accompanying paper](https://stevendiamond.me/pdf/proximal.pdf) for a fu
 
 A tutorial and other documentation can be found at [proximal-lang.org](http://www.proximal-lang.org/).
 
-This git repository holds the latest development version of ProxImaL. For installation instructions, 
+This git repository holds the latest development version of ProxImaL. For installation instructions,
 see the [install guide](http://www.proximal-lang.org/en/latest/install/index.html) at [proximal-lang.org](http://www.proximal-lang.org/).

doc/source/index.rst

@@ -6,8 +6,8 @@
 Welcome to ProxImaL
 =====================
 
-ProxImaL is a Python-embedded modeling language for image optimization problems. 
-It allows you to express your problem in a natural way that follows the math, 
+ProxImaL is a Python-embedded modeling language for image optimization problems.
+It allows you to express your problem in a natural way that follows the math,
 and automatically determines an efficient method for solving the problem.
 ProxImaL makes it easy to experiment with many different priors and other problem reformulations,
 without worrying about the details of how the problem is solved.
@@ -22,7 +22,7 @@ For example, the following code denoises an image using simple sparse gradient a
           import matplotlib.pyplot as plt
 
           # Generate data.
-          I = scipy.misc.ascent()
+          I = scipy.datasets.ascent()
           np.random.seed(1)
           b = I + 10*np.random.randn(*I.shape)
 

proximal/examples/basic_denoise.py

@@ -9,7 +9,7 @@
 from proximal.utils.utils import Impl
 
 # Generate data.
-I = scipy.misc.ascent()
+I = scipy.datasets.ascent()
 np.random.seed(1)
 b = I + 10 * np.random.randn(*I.shape)
 

proximal/examples/test_mask.py

@@ -47,7 +47,7 @@
 
 # Error
 print('Maximum error {0}'.format(
-    np.linalg.norm(output_ref.ravel() - output.ravel(), np.Inf)))
+    np.linalg.norm(output_ref.ravel() - output.ravel(), np.inf)))
 
 plt.subplot(223)
 plt.imshow(output_ref, interpolation="nearest", clim=(0.0, 1.0), cmap='gray')
@@ -73,6 +73,6 @@
 
 # Error
 print('Maximum error {0}'.format(
-    np.linalg.norm(output_ref.ravel() - output_trans.ravel(), np.Inf)))
+    np.linalg.norm(output_ref.ravel() - output_trans.ravel(), np.inf)))
 
 plt.show()

proximal/examples/test_warp.py

@@ -70,7 +70,7 @@
 plt.title('Output from halide')
 
 # Error
-delta = np.linalg.norm(output_ref.ravel() - output.ravel(), np.Inf)
+delta = np.linalg.norm(output_ref.ravel() - output.ravel(), np.inf)
 norm = np.amax((output_ref.max(), output.max()))
 print('Relative error {0}'.format(delta / norm))
 
@@ -111,7 +111,7 @@
 plt.title('Output trans from CV2')
 
 # Error
-delta = np.linalg.norm(output_ref_trans.ravel() - output_trans.ravel(), np.Inf)
+delta = np.linalg.norm(output_ref_trans.ravel() - output_trans.ravel(), np.inf)
 norm = np.amax((output_ref_trans.max(), output_trans.max()))
 print('Relative error trans {0}'.format(delta / norm))
 plt.show()

proximal/tests/base_test.py

@@ -20,7 +20,7 @@ def assertItemsAlmostEqual(self,
 
         assert a.shape == b.shape
 
-        norm_infinity_b = np.linalg.norm(b.ravel(), np.Inf)
+        norm_infinity_b = np.linalg.norm(b.ravel(), np.inf)
 
         if norm_infinity_b == 0:
             # b is a zero matrix / vector

proximal/tests/test_halide.py

@@ -1,6 +1,5 @@
-import os
 import numpy as np
-from scipy.misc import ascent
+from scipy.datasets import ascent
 from scipy.signal import convolve2d
 
 import proximal as px
@@ -17,7 +16,7 @@ def test_compile(self):
         """Test compilation
         """
         Halide('A_conv', reconfigure=True, recompile=True, verbose=False)
-    
+
 
     def _get_testvector(self):
         # Load image
@@ -28,7 +27,7 @@ def _get_testvector(self):
         K /= K.size
 
         return np_img, K
-    
+
 
     def test_run(self):
         """ Test running
@@ -43,7 +42,7 @@ def test_run(self):
         output_ref = convolve2d(np_img, K, mode='same', boundary='wrap')
 
         self.assertItemsAlmostEqual(output, output_ref)
-    
+
     def _test_algo(self, algo, check_convergence=True):
         """ Ensure all internal buffers of the algortihms are Fortran-style ordered.
         """

proximal/utils/utils.py

@@ -2,7 +2,7 @@
 
 # Imports
 import numpy as np
-from scipy.misc import ascent, face
+from scipy.datasets import ascent, face
 import cv2
 import os
 from numpy.fft import fftn, ifftn, fft2, ifft2

pyproject.toml

@@ -26,6 +26,9 @@ test = [
     "pytest",
     "pytest-cov",
     "ruff",
+
+    # Required by scipy.datasets.accent()
+    "pooch",
 ]
 
 [project.urls]