test: adding pytest.parametrize and numpy.polynomial

Author: Luiskitsu

news/test_squeeze.rst

@@ -1,6 +1,6 @@
 **Added:**
 
-* Squeeze test: given a squeezed signal, unsqueezing recovers the original
+* Polynomial squeeze of x-axis of morphed data
 
 **Changed:**
 

tests/test_morphsqueeze.py

@@ -1,54 +1,40 @@
-"""
-The squeeze morph is used to correct for small non-linear geometric distortions
-from detectors that are not effectively corrected during calibration, such as
-individual module misalignment or tilt. This squeezing is applied as:
-x_squeezed = x + squeeze_0 + squeeze_1 * x**2 + squeeze_2 * x**3
-
-The squeeze distortions that we might encounter practically are going to be
-very small. Furthermore, large values for the squeezing parameters lead to
-missing values during interpolation. Therefore is important to use small
-squeezing values to avoid error or unphysical results.
-Safe bounds for the squeezing parameters are:
-squeeze_0 [-0.2, 0.2]
-squeeze_1 [-0.001, 0.001]
-squeeze_2 [-0.0001, 0.0001]
-Values outside these bounds should be used carefully.
-Note that these bounds are established for an x-axis that goes from 0 to 10.
-"""
-
-import matplotlib.pyplot as plt
 import numpy as np
+import pytest
+from numpy.polynomial import Polynomial
 from scipy.interpolate import interp1d
 
 
-def test_morphsqueeze():
-    """
-    Test that we can unsqueeze squeezed data.
-    The test inputs are an expected uniform target (e.g. synchrotron data)
-    and a squeezed version of the target (e.g. XFEL data). The squeezed data
-    is created by applying a nonlinear distortion to the uniform target.
-    Both input data are in the same uniform x-axis grid.
-    Then we unsqueeze the squeezed data by doing the inverse transformation
-    using interpolatiion.
-    Finally we check that the unsqueezed data matches the expected uniform
-    target.
-    """
-
+@pytest.mark.parametrize(
+    "squeeze_coeffs",
+    [
+        # The order of coefficients is [a0, a1, a2, ..., an]
+        # Negative cubic squeeze coefficients
+        [-0.2, -0.01, -0.001, -0.001],
+        # Positive cubic squeeze coefficients
+        [0.2, 0.01, 0.001, 0.001],
+        # Positive and negative cubic squeeze coefficients
+        [0.2, -0.01, 0.001, -0.001],
+        # Quadratic squeeze coefficients
+        [-0.2, 0.005, -0.003],
+        # Linear squeeze coefficients
+        [0.1, 0.3],
+        # 4th order squeeze coefficients
+        [0.2, -0.01, 0.001, -0.001, 0.0001],
+    ],
+)
+def test_morphsqueeze(squeeze_coeffs):
     # Uniform x-axis grid. This is the same x-axis for all data.
     x = np.linspace(0, 10, 1000)
     # Expected uniform target
     y_expected = np.sin(x)
 
+    # Create polynomial based on a list of values for polynomial coefficients
+    squeeze_polynomial = Polynomial(squeeze_coeffs)
     # Apply squeeze parameters to uniform data to get the squeezed data
-    # Include squeeze_0 for squeezes with offset
-    squeeze_0 = 0.2
-    squeeze_1 = -0.001
-    squeeze_2 = -0.001
-    x_squeezed = x + squeeze_0 + squeeze_1 * x**2 + squeeze_2 * x**3
+    x_squeezed = x + squeeze_polynomial(x)
     y_squeezed = np.sin(x_squeezed)
 
     # Unsqueeze the data by interpolating back to uniform grid
-    # y_unsqueezed = np.interp(x, x_squeezed, y_squeezed)
     y_unsqueezed = interp1d(
         x_squeezed,
         y_squeezed,
@@ -59,16 +45,16 @@ def test_morphsqueeze():
     y_actual = y_unsqueezed
 
     # Check that the unsqueezed (actual) data matches the expected data
-    # Including tolerance error because I was having issues
-    # with y_actual == y_expected. I think is because interpolation?
-    assert np.allclose(y_actual, y_expected, atol=100)
-
-    # Plot to verify what we are doing
-    plt.figure(figsize=(7, 4))
-    plt.plot(x, y_expected, color="black", label="Expected uniform data")
-    plt.plot(x, y_squeezed, "--", color="purple", label="Squeezed data")
-    plt.plot(x, y_unsqueezed, "--", color="gold", label="Unsqueezed data")
-    plt.xlabel("x")
-    plt.ylabel("y")
-    plt.legend()
-    plt.show()
+    # Including tolerance error because of extrapolation error
+    assert np.allclose(y_actual, y_expected, atol=1)
+
+    # This plotting code was used for the comments in the github
+    # PR https://github.com/diffpy/diffpy.morph/pull/180
+    # plt.figure(figsize=(7, 4))
+    # plt.plot(x, y_expected, color="black", label="Expected uniform data")
+    # plt.plot(x, y_squeezed, "--", color="purple", label="Squeezed data")
+    # plt.plot(x, y_unsqueezed, "--", color="gold", label="Unsqueezed data")
+    # plt.xlabel("x")
+    # plt.ylabel("y")
+    # plt.legend()
+    # plt.show()