Updated test file to use pytest as well.

tests/test_implementation.py

@@ -19,6 +19,9 @@
 from __future__ import unicode_literals
 from __future__ import absolute_import
 
+import unittest
+
+import pytest
 import six
 import numpy as np
 from numpy.testing import assert_allclose
@@ -28,40 +31,36 @@
 from .lspopt_ref import lspopt_ref
 
 
-class TestLSPOptSuite(object):
-
-    def test_different_N(self):
-        """Test against reference implementation for different N."""
-        def test_fcn(n):
-            h1, w1 = lspopt(n)
-            h2, w2 = lspopt_ref(n)
-            assert_allclose(h1, h2)
-            assert_allclose(w1, w2)
+@pytest.mark.parametrize("n", six.moves.range(64, 1024))
+def test_different_n(n):
+    """Test against reference implementation for different N."""
+    h1, w1 = lspopt(n)
+    h2, w2 = lspopt_ref(n)
+    assert_allclose(h1, h2)
+    assert_allclose(w1, w2)
 
-        for n in six.moves.range(64, 1024):
-            yield test_fcn, n
 
-    def test_different_c(self):
-        """Test against reference implementation for different N."""
-        def test_fcn(c):
-            h1, w1 = lspopt(n=1024, c_parameter=c)
-            h2, w2 = lspopt_ref(n=1024, c_parameter=c)
-            assert_allclose(h1, h2)
-            assert_allclose(w1, w2)
+@pytest.mark.parametrize("c", np.arange(1.1, 30.0, 0.1))
+def test_different_c(c):
+    """Test against reference implementation for different c."""
+    h1, w1 = lspopt(n=1024, c_parameter=c)
+    h2, w2 = lspopt_ref(n=1024, c_parameter=c)
+    assert_allclose(h1, h2)
+    assert_allclose(w1, w2)
 
-        for c in np.arange(1.1, 30.0, 0.1):
-            yield test_fcn, c
 
-    def test_spectrogram_method(self):
+def test_spectrogram_method():
+    """Test the spectrogram method's functionality."""
+    fs = 10e3
+    N = 1e5
+    amp = 2 * np.sqrt(2)
+    noise_power = 0.001 * fs / 2
+    time = np.arange(N) / fs
+    freq = np.linspace(1e3, 2e3, N)
+    x = amp * chirp(time, 1e3, 2.0, 6e3, method='quadratic') + \
+        np.random.normal(scale=np.sqrt(noise_power), size=time.shape)
 
-        fs = 10e3
-        N = 1e5
-        amp = 2 * np.sqrt(2)
-        noise_power = 0.001 * fs / 2
-        time = np.arange(N) / fs
-        freq = np.linspace(1e3, 2e3, N)
-        x = amp * chirp(time, 1e3, 2.0, 6e3, method='quadratic') + \
-            np.random.normal(scale=np.sqrt(noise_power), size=time.shape)
+    f, t, Sxx = spectrogram_lspopt(x, fs, c_parameter=20.0)
+    f_sp, t_sp, Sxx_sp = spectrogram(x, fs)
 
-        f, t, Sxx = spectrogram_lspopt(x, fs, c_parameter=20.0)
-        f_sp, t_sp, Sxx_sp = spectrogram(x, fs)
+    assert True