Merge 09cae1d into dd4b46b

dev-environment.yml

@@ -13,7 +13,7 @@ dependencies:
   - tqdm
   - scikit-image=0.*
   - scikit-gstat>=1.0
-  - geoutils>=0.1.4,<0.2
+  - geoutils>=0.1.5,<0.2
 
   # Development-specific, to mirror manually in setup.cfg [options.extras_require].
   - pip
@@ -51,4 +51,4 @@ dependencies:
     - opencv-contrib-python-headless
 
     # To run CI against latest GeoUtils
-#    - git+https://github.com/rhugonnet/geoutils.git@fix_to_points
+#    - git+https://github.com/rhugonnet/geoutils.git

doc/source/code/comparison_plot_spatial_interpolation.py

@@ -13,7 +13,7 @@
 # The example DEMs are void-free, so let's make some random voids.
 ddem.data.mask = np.zeros_like(ddem.data, dtype=bool)  # Reset the mask
 # Introduce 50000 nans randomly throughout the dDEM.
-ddem.data.mask.ravel()[np.random.choice(ddem.data.size, 50000, replace=False)] = True
+ddem.data.mask.ravel()[np.random.default_rng(42).choice(ddem.data.size, 50000, replace=False)] = True
 
 ddem.interpolate(method="linear")
 

doc/source/code/spatialstats_standardizing.py

@@ -5,8 +5,8 @@
 # Example x vector
 mu = 15
 sig = 5
-np.random.seed(42)
-y = np.random.normal(mu, sig, size=300)
+rng = np.random.default_rng(42)
+y = rng.normal(mu, sig, size=300)
 
 fig, ax1 = plt.subplots(figsize=(8, 3))
 

doc/source/code/spatialstats_stationarity_assumption.py

@@ -8,10 +8,10 @@
 x = np.linspace(0, 1, 200)
 
 sig = 0.2
-np.random.seed(42)
-y_rand1 = np.random.normal(0, sig, size=len(x))
-y_rand2 = np.random.normal(0, sig, size=len(x))
-y_rand3 = np.random.normal(0, sig, size=len(x))
+rng = np.random.default_rng(42)
+y_rand1 = rng.normal(0, sig, size=len(x))
+y_rand2 = rng.normal(0, sig, size=len(x))
+y_rand3 = rng.normal(0, sig, size=len(x))
 
 
 y_mean = np.array([0.5 * xval - 0.25 if xval > 0.5 else 0.5 * (1 - xval) - 0.25 for xval in x])

environment.yml

@@ -13,9 +13,9 @@ dependencies:
   - tqdm
   - scikit-image=0.*
   - scikit-gstat>=1.0
-  - geoutils>=0.1.4,<0.2
+  - geoutils>=0.1.5,<0.2
   - pip
 
    # To run CI against latest GeoUtils
 #  - pip:
-#    - git+https://github.com/rhugonnet/geoutils.git@fix_to_points
+#    - git+https://github.com/rhugonnet/geoutils.git

examples/advanced/plot_norm_regional_hypso.py

@@ -58,8 +58,9 @@
 # %%
 # To test the method, we can generate a semi-random mask to assign nans to glacierized areas.
 # Let's remove 30% of the data.
-np.random.seed(42)
-random_nans = (xdem.misc.generate_random_field(dem_1990.shape, corr_size=5) > 0.7) & (glacier_index_map > 0)
+random_nans = (xdem.misc.generate_random_field(dem_1990.shape, corr_size=5, random_state=42) > 0.7) & (
+    glacier_index_map > 0
+)
 
 random_nans.plot()
 

requirements.txt

@@ -11,5 +11,5 @@ scipy>=1.0,<1.13
 tqdm
 scikit-image==0.*
 scikit-gstat>=1.0
-geoutils>=0.1.4,<0.2
+geoutils>=0.1.5,<0.2
 pip

tests/test_coreg/test_affine.py

@@ -143,7 +143,7 @@ def test_coreg_example(self, verbose: bool = False) -> None:
 
         # Check the output metadata is always the same
         shifts = (nuth_kaab._meta["offset_east_px"], nuth_kaab._meta["offset_north_px"], nuth_kaab._meta["vshift"])
-        assert shifts == pytest.approx((-0.463, -0.133, -1.9876264671765433))
+        assert shifts == pytest.approx((-0.463, -0.1339999, -1.9922009))
 
     def test_gradientdescending(self, subsample: int = 10000, inlier_mask: bool = True, verbose: bool = False) -> None:
         """

tests/test_coreg/test_base.py

@@ -95,7 +95,8 @@ def test_error_method(self) -> None:
         assert vshiftcorr.error(dem1, dem2, transform=affine, crs=crs, error_type="median") == -2
 
         # Create random noise and see if the standard deviation is equal (it should)
-        dem3 = dem1.copy() + np.random.random(size=dem1.size).reshape(dem1.shape)
+        rng = np.random.default_rng(42)
+        dem3 = dem1.copy() + rng.random(size=dem1.size).reshape(dem1.shape)
         assert abs(vshiftcorr.error(dem1, dem3, transform=affine, crs=crs, error_type="std") - np.std(dem3)) < 1e-6
 
     @pytest.mark.parametrize("subsample", [10, 10000, 0.5, 1])  # type: ignore
@@ -104,8 +105,8 @@ def test_get_subsample_on_valid_mask(self, subsample: float | int) -> None:
 
         # Define a valid mask
         width = height = 50
-        np.random.seed(42)
-        valid_mask = np.random.randint(low=0, high=2, size=(width, height), dtype=bool)
+        rng = np.random.default_rng(42)
+        valid_mask = rng.integers(low=0, high=2, size=(width, height), dtype=bool)
 
         # Define a class with a subsample and random_state in the metadata
         coreg = Coreg(meta={"subsample": subsample, "random_state": 42})

tests/test_coreg/test_biascorr.py

@@ -42,7 +42,6 @@ class TestBiasCorr:
     )
 
     # Convert DEMs to points with a bit of subsampling for speed-up
-    # TODO: Simplify once this GeoUtils issue is resolved: https://github.com/GlacioHack/geoutils/issues/499
     tba_pts = tba.to_pointcloud(data_column_name="z", subsample=50000, random_state=42).ds
 
     ref_pts = ref.to_pointcloud(data_column_name="z", subsample=50000, random_state=42).ds
@@ -328,6 +327,8 @@ def test_biascorr__bin_and_fit_1d(self, fit_args, fit_func, fit_optimizer, bin_s
         # Curve fit can be unhappy in certain circumstances for numerical estimation of covariance
         # We don't care for this test
         warnings.filterwarnings("ignore", message="Covariance of the parameters could not be estimated*")
+        # Apply the transform can create data exactly equal to the nodata
+        warnings.filterwarnings("ignore", category=UserWarning, message="Unmasked values equal to the nodata value*")
 
         # Create a bias correction object
         bcorr = biascorr.BiasCorr(
@@ -421,7 +422,6 @@ def test_directionalbias__synthetic(self, fit_args, angle, nb_freq) -> None:
         xx = gu.raster.get_xy_rotated(self.ref, along_track_angle=angle)[0]
 
         # Get random parameters (3 parameters needed per frequency)
-        np.random.seed(42)
         params = np.array([(5, 3000, np.pi), (1, 300, 0), (0.5, 100, np.pi / 2)]).flatten()
         nb_freq = 1
         params = params[0 : 3 * nb_freq]
@@ -510,8 +510,8 @@ def test_deramp__synthetic(self, fit_args, order: int) -> None:
         nb_params = int((order + 1) * (order + 1))
 
         # Get a random number of parameters
-        np.random.seed(42)
-        params = np.random.normal(size=nb_params)
+        rng = np.random.default_rng(42)
+        params = rng.normal(size=nb_params)
 
         # Create a synthetic bias and add to the DEM
         synthetic_bias = polynomial_2d((xx, yy), *params)

tests/test_ddem.py

@@ -52,7 +52,8 @@ def test_regional_hypso(self) -> None:
         """Test the regional hypsometric approach."""
         ddem = self.ddem.copy()
         ddem.data.mask = np.zeros_like(ddem.data, dtype=bool)
-        ddem.data.mask.ravel()[np.random.choice(ddem.data.size, 50000, replace=False)] = True
+        rng = np.random.default_rng(42)
+        ddem.data.mask.ravel()[rng.choice(ddem.data.size, 50000, replace=False)] = True
         assert np.count_nonzero(ddem.data.mask) > 0
 
         assert ddem.filled_data is None
@@ -71,7 +72,8 @@ def test_local_hypso(self) -> None:
         ddem = self.ddem.copy()
         scott_1990 = self.outlines_1990.query("NAME == 'Scott Turnerbreen'")
         ddem.data.mask = np.zeros_like(ddem.data, dtype=bool)
-        ddem.data.mask.ravel()[np.random.choice(ddem.data.size, 50000, replace=False)] = True
+        rng = np.random.default_rng(42)
+        ddem.data.mask.ravel()[rng.choice(ddem.data.size, 50000, replace=False)] = True
         assert np.count_nonzero(ddem.data.mask) > 0
 
         assert ddem.filled_data is None

tests/test_demcollection.py

@@ -60,10 +60,11 @@ def test_init(self) -> None:
         # Simple check that the dV number is of a greater magnitude than the dH number.
         assert abs(cumulative_dv.iloc[-1]) > abs(cumulative_dh.iloc[-1])
 
+        rng = np.random.default_rng(42)
         # Generate 10000 NaN values randomly in one of the dDEMs
         dems.ddems[0].data[
-            np.random.randint(0, dems.ddems[0].data.shape[0], 100),
-            np.random.randint(0, dems.ddems[0].data.shape[1], 100),
+            rng.integers(0, dems.ddems[0].data.shape[0], 100),
+            rng.integers(0, dems.ddems[0].data.shape[1], 100),
         ] = np.nan
         # Check that the cumulative_dh function warns for NaNs
         with warnings.catch_warnings():
@@ -108,9 +109,10 @@ def test_ddem_interpolation(self) -> None:
                 raise exception
 
         # Generate 10000 NaN values randomly in one of the dDEMs
+        rng = np.random.default_rng(42)
         dems.ddems[0].data[
-            np.random.randint(0, dems.ddems[0].data.shape[0], 100),
-            np.random.randint(0, dems.ddems[0].data.shape[1], 100),
+            rng.integers(0, dems.ddems[0].data.shape[0], 100),
+            rng.integers(0, dems.ddems[0].data.shape[1], 100),
         ] = np.nan
 
         # Make sure that filled_data is not available anymore, since the data now has nans

tests/test_examples.py

@@ -28,17 +28,17 @@ class TestExamples:
     @pytest.mark.parametrize(
         "rst_and_truevals",
         [
-            (ref_dem, np.array([868.6489, 623.42194, 180.57921, 267.30765, 601.67615], dtype=np.float32)),
-            (tba_dem, np.array([875.2358, 625.0544, 182.9936, 272.6586, 606.2897], dtype=np.float32)),
+            (ref_dem, np.array([465.11816, 207.3236, 208.30563, 748.7337, 797.28644], dtype=np.float32)),
+            (tba_dem, np.array([464.6715, 213.7554, 207.8788, 760.8192, 797.3268], dtype=np.float32)),
             (
                 ddem,
                 np.array(
                     [
-                        -0.012023926,
-                        -0.6956787,
-                        0.14024353,
-                        1.1026001,
-                        -5.9224243,
+                        1.3182373,
+                        -1.6629944,
+                        0.10473633,
+                        -10.096802,
+                        2.4724731,
                     ],
                     dtype=np.float32,
                 ),
@@ -50,8 +50,8 @@ def test_array_content(self, rst_and_truevals: tuple[Raster, NDArrayf]) -> None:
 
         rst = rst_and_truevals[0]
         truevals = rst_and_truevals[1]
-        np.random.seed(42)
-        values = np.random.choice(rst.data.data.flatten(), size=5, replace=False)
+        rng = np.random.default_rng(42)
+        values = rng.choice(rst.data.data.flatten(), size=5, replace=False)
 
         assert values == pytest.approx(truevals)
 

tests/test_filters.py

@@ -37,8 +37,9 @@ def test_gauss(self) -> None:
 
         # Test that it works with NaNs too
         nan_count = 1000
-        cols = np.random.randint(0, high=self.dem_1990.width - 1, size=nan_count, dtype=int)
-        rows = np.random.randint(0, high=self.dem_1990.height - 1, size=nan_count, dtype=int)
+        rng = np.random.default_rng(42)
+        cols = rng.integers(0, high=self.dem_1990.width - 1, size=nan_count, dtype=int)
+        rows = rng.integers(0, high=self.dem_1990.height - 1, size=nan_count, dtype=int)
         dem_with_nans = np.copy(self.dem_1990.data).squeeze()
         dem_with_nans[rows, cols] = np.nan
 
@@ -71,8 +72,9 @@ def test_dist_filter(self) -> None:
 
         # Add random outliers
         count = 1000
-        cols = np.random.randint(0, high=self.dem_1990.width - 1, size=count, dtype=int)
-        rows = np.random.randint(0, high=self.dem_1990.height - 1, size=count, dtype=int)
+        rng = np.random.default_rng(42)
+        cols = rng.integers(0, high=self.dem_1990.width - 1, size=count, dtype=int)
+        rows = rng.integers(0, high=self.dem_1990.height - 1, size=count, dtype=int)
         ddem.data[rows, cols] = 5000
 
         # Filter gross outliers

tests/test_fit.py

@@ -53,15 +53,15 @@ def test_robust_norder_polynomial_fit_noise_and_outliers(self) -> None:
         # Ignore sklearn convergence warnings
         warnings.filterwarnings("ignore", category=UserWarning, message="lbfgs failed to converge")
 
-        np.random.seed(42)
+        rng = np.random.default_rng(42)
 
         # Define x vector
         x = np.linspace(1, 10, 1000)
         # Define an exact polynomial
         true_coefs = [-100, 5, 3, 2]
         y = np.polyval(np.flip(true_coefs), x).astype(np.float32)
         # Add some noise on top
-        y += np.random.normal(loc=0, scale=3, size=1000)
+        y += rng.normal(loc=0, scale=3, size=1000)
         # Add some outliers
         y[50:75] = 0.0
         y[900:925] = 1000.0
@@ -103,8 +103,8 @@ def test_robust_norder_polynomial_fit_noise_and_outliers(self) -> None:
         coefs4, deg4 = xdem.fit.robust_norder_polynomial_fit(x, y, estimator_name="Theil-Sen", random_state=42)
         assert deg4 == 3
         # High degree coefficients should be well constrained
-        assert coefs4[2] == pytest.approx(true_coefs[2], abs=1)
-        assert coefs4[3] == pytest.approx(true_coefs[3], abs=1)
+        assert coefs4[2] == pytest.approx(true_coefs[2], abs=1.5)
+        assert coefs4[3] == pytest.approx(true_coefs[3], abs=1.5)
 
         # RANSAC also works
         coefs5, deg5 = xdem.fit.robust_norder_polynomial_fit(x, y, estimator_name="RANSAC", random_state=42)
@@ -150,15 +150,15 @@ def test_robust_nfreq_sumsin_fit(self) -> None:
     def test_robust_nfreq_simsin_fit_noise_and_outliers(self) -> None:
 
         # Check robustness to outliers
-        np.random.seed(42)
+        rng = np.random.default_rng(42)
         # Define X vector
         x = np.linspace(0, 10, 1000)
         # Define exact sum of sinusoid signal
         true_coefs = np.array([(5, 3, np.pi), (3, 0.5, 0)]).flatten()
         y = xdem.fit.sumsin_1d(x, *true_coefs)
 
         # Add some noise
-        y += np.random.normal(loc=0, scale=0.25, size=1000)
+        y += rng.normal(loc=0, scale=0.25, size=1000)
         # Add some outliers
         y[50:75] = -10
         y[900:925] = 10