Merge fe10752 into 9a09cd1

MANIFEST.in

@@ -1,4 +1,3 @@
 include docs/source/*.rst
 include docs/source/examples/*.txt
 include docs/source/examples/*.py
-include gpfit/tests/artifacts/*.txt

gpfit/tests/README.md

@@ -0,0 +1,10 @@
+## Plot unit tests
+
+Plot unit tests are run with [pytest-mpl](https://github.com/matplotlib/pytest-mpl)
+
+To generate the baseline images:
+`pytest --mpl-generate-path=baseline t_plot_fit.py`
+
+To test against the baseline images:
+`pytest --mpl t_plot_fit.py`
+

gpfit/tests/__init__.py

@@ -1,7 +0,0 @@
-"""
-To run all GPfit unit tests, execute:
-
-from gpfit.tests import run
-run()
-"""
-from .run_tests import run

gpfit/tests/run_tests.sh

@@ -0,0 +1 @@
+pytest --junitxml="test_reports" --mpl

gpfit/tests/t_fit.py → gpfit/tests/test_fit.py

@@ -78,18 +78,18 @@ def test_incorrect_inputs(self):
 
     def test_save_and_load(self):
         f1 = ImplicitSoftmaxAffine(self.x, self.y, self.K, seed=SEED)
-        f1.save("artifacts/fit.pkl")
+        f1.save("fit.pkl")
         strings1 = f1.__repr__()
-        with open("artifacts/fit.pkl", "rb") as picklefile:
+        with open("fit.pkl", "rb") as picklefile:
             f2 = pickle.load(picklefile)
         self.assertTrue(f2.errors["rms_rel"] < 1e-5)
         strings2 = f2.__repr__()
         self.assertEqual(strings1, strings2)
 
     def test_savetxt(self):
         f = ImplicitSoftmaxAffine(self.x, self.y, self.K, seed=SEED)
-        f.savetxt("artifacts/fit.txt")
-        with open("artifacts/fit.txt", "r", encoding="utf-8") as textfile:
+        f.savetxt("fit.txt")
+        with open("fit.txt", "r", encoding="utf-8") as textfile:
             fitstring = textfile.read()
         self.assertEqual(fitstring, (
             "1 = (0.947385/w**0.0920329)*(u_1)**0.0176859\n"

gpfit/tests/t_plot_fit.py → gpfit/tests/test_plot_fit.py

@@ -1,10 +1,12 @@
 "unit tests for plot fit methods"
-import unittest
+import pytest
 import numpy as np
 from gpfit.fit import MaxAffine, SoftmaxAffine, ImplicitSoftmaxAffine
 
+SEED = 33404
 
-class TestPlot(unittest.TestCase):
+
+class TestPlot:
     "Unit tests for plot methods"
 
     N = 51
@@ -14,23 +16,26 @@ class TestPlot(unittest.TestCase):
     y = np.log(w)
     K = 2
 
+    @pytest.mark.mpl_image_compare(filename='ma_1d.png')
     def test_max_affine(self):
-        f = MaxAffine(self.x, self.y, self.K)
+        f = MaxAffine(self.x, self.y, self.K, seed=SEED)
         fig, _ = f.plot()
-        fig.savefig("artifacts/ma_test.png")
+        return fig
 
+    @pytest.mark.mpl_image_compare(filename='sma_1d.png')
     def test_softmax_affine(self):
-        f = SoftmaxAffine(self.x, self.y, self.K)
+        f = SoftmaxAffine(self.x, self.y, self.K, seed=SEED)
         fig, _ = f.plot()
-        fig.savefig("artifacts/sma_test.png")
+        return fig
 
+    @pytest.mark.mpl_image_compare(filename='isma_1d.png')
     def test_implicit_softmax_affine(self):
-        f = ImplicitSoftmaxAffine(self.x, self.y, self.K)
+        f = ImplicitSoftmaxAffine(self.x, self.y, self.K, seed=SEED)
         fig, _ = f.plot()
-        fig.savefig("artifacts/isma_test.png")
+        return fig
 
 
-class TestPlotSurface(unittest.TestCase):
+class TestPlotSurface:
     "Unit tests for plot_surface methods"
 
     rng = np.random.RandomState(33404)
@@ -42,23 +47,26 @@ class TestPlotSurface(unittest.TestCase):
     y = np.log(P)
     K = 3
 
+    @pytest.mark.mpl_image_compare(filename='ma_2d_surface.png')
     def test_max_affine(self):
-        f = MaxAffine(self.x, self.y, self.K)
+        f = MaxAffine(self.x, self.y, self.K, seed=SEED)
         fig, _ = f.plot_surface(azim=135)
-        fig.savefig("artifacts/ma_test_surface.png")
+        return fig
 
+    @pytest.mark.mpl_image_compare(filename='sma_2d_surface.png')
     def test_softmax_affine(self):
-        f = SoftmaxAffine(self.x, self.y, self.K)
+        f = SoftmaxAffine(self.x, self.y, self.K, seed=SEED)
         fig, _ = f.plot_surface(azim=135)
-        fig.savefig("artifacts/sma_test_surface.png")
+        return fig
 
+    @pytest.mark.mpl_image_compare(filename='isma_2d_surface.png')
     def test_implicit_softmax_affine(self):
-        f = ImplicitSoftmaxAffine(self.x, self.y, self.K)
+        f = ImplicitSoftmaxAffine(self.x, self.y, self.K, seed=SEED)
         fig, _ = f.plot_surface(azim=135)
-        fig.savefig("artifacts/isma_test_surface.png")
+        return fig
 
 
-class TestPlotSlices(unittest.TestCase):
+class TestPlotSlices:
     "Unit tests for plot_slices method"
 
     Vdd = np.linspace(1, 2, 10)
@@ -71,33 +79,26 @@ class TestPlotSlices(unittest.TestCase):
     y = np.log(P)
     K = 3
 
+    @pytest.mark.mpl_image_compare(filename='ma_2d_slices.png')
     def test_max_affine(self):
-        f = MaxAffine(self.x, self.y, self.K)
+        f = MaxAffine(self.x, self.y, self.K, seed=SEED)
         fig, _ = f.plot_slices()
-        fig.savefig("artifacts/ma_test_slices.png")
+        return fig
 
+    @pytest.mark.mpl_image_compare(filename='sma_2d_slices.png')
     def test_softmax_affine(self):
-        f = SoftmaxAffine(self.x, self.y, self.K)
+        f = SoftmaxAffine(self.x, self.y, self.K, seed=SEED)
         fig, _ = f.plot_slices()
-        fig.savefig("artifacts/sma_test_slices.png")
+        return fig
 
+    @pytest.mark.mpl_image_compare(filename='isma_2d_slices.png')
     def test_implicit_softmax_affine(self):
-        f = ImplicitSoftmaxAffine(self.x, self.y, self.K)
+        f = ImplicitSoftmaxAffine(self.x, self.y, self.K, seed=SEED)
         fig, _ = f.plot_slices()
-        fig.savefig("artifacts/isma_test_slices.png")
-
+        return fig
 
 TESTS = [
     TestPlot,
     TestPlotSurface,
     TestPlotSlices,
 ]
-
-if __name__ == '__main__':
-    SUITE = unittest.TestSuite()
-    LOADER = unittest.TestLoader()
-
-    for t in TESTS:
-        SUITE.addTests(LOADER.loadTestsFromTestCase(t))
-
-    unittest.TextTestRunner(verbosity=2).run(SUITE)

setup.py

@@ -30,9 +30,17 @@
     author_email="gpkit@mit.edu",
     url="https://github.com/convexengineering/gpfit",
     python_requires=">=3.4",
-    install_requires=["numpy", "scipy", "gpkit", "matplotlib"],
+    install_requires=[
+        "numpy",
+        "scipy",
+        "gpkit",
+        "matplotlib==3.3.2",
+        "pytest==6.2.5",
+        "pytest-mpl==0.13",
+    ],
     version="0.2.0",
-    packages=["gpfit", "gpfit.maths", "gpfit.tests", "gpfit.xfoil"],
+    packages=["gpfit", "gpfit.maths", "gpfit.tests", "gpfit.tests.baseline",
+              "gpfit.xfoil"],
     include_package_data=True,
     license=LICENSE,
 )