simplify tension tests and add test for direct input of nested sampli…

…ng stats to tension stats

tests/test_tension.py

@@ -1,7 +1,8 @@
-from pytest import approx
+from pytest import approx, raises
 from anesthetic.examples.perfect_ns import correlated_gaussian
 import numpy as np
 from numpy.linalg import inv, slogdet
+from pandas.testing import assert_series_equal
 from anesthetic.tension import stats
 
 
@@ -37,30 +38,32 @@ def test_tension_stats_compatible_gaussian():
 
     nsamples = 10
     beta = 1
-    samples_stats = stats(samplesA, samplesB, samplesAB, nsamples, beta)
+    s = stats(samplesA, samplesB, samplesAB, nsamples, beta)
 
-    logR_std = samples_stats.logR.std()
-    logR_mean = samples_stats.logR.mean()
     logR_exact = logV - dmu_cov_dmu_AB/2 - slogdet(2*np.pi*(covA+covB))[1]/2
-    assert logR_mean == approx(logR_exact, abs=3*logR_std)
+    assert s.logR.mean() == approx(logR_exact, abs=3*s.logR.std())
 
-    logS_std = samples_stats.logS.std()
-    logS_mean = samples_stats.logS.mean()
     logS_exact = d / 2 - dmu_cov_dmu_AB / 2
-    assert logS_mean == approx(logS_exact, abs=3*logS_std)
+    assert s.logS.mean() == approx(logS_exact, abs=3*s.logS.std())
 
-    logI_std = samples_stats.logI.std()
-    logI_mean = samples_stats.logI.mean()
     logI_exact = logV - d / 2 - slogdet(2*np.pi*(covA+covB))[1] / 2
-    assert logI_mean == approx(logI_exact, abs=3*logI_std)
+    assert s.logI.mean() == approx(logI_exact, abs=3*s.logI.std())
 
-    assert logS_mean == approx(logR_mean - logI_mean, abs=3*logS_std)
+    assert s.logS.mean() == approx(s.logR.mean() - s.logI.mean(),
+                                   abs=3*s.logS.std())
 
-    assert samples_stats.get_labels().tolist() == ([r'$\ln\mathcal{R}$',
-                                                    r'$\ln\mathcal{I}$',
-                                                    r'$\ln\mathcal{S}$',
-                                                    r'$d_\mathrm{G}$',
-                                                    r'$p$'])
+    assert s.get_labels().tolist() == ([r'$\ln\mathcal{R}$',
+                                        r'$\ln\mathcal{I}$',
+                                        r'$\ln\mathcal{S}$',
+                                        r'$d_\mathrm{G}$',
+                                        r'$p$'])
+
+    with raises(ValueError):
+        stats(samplesA.stats(nsamples=5), samplesB, samplesAB, nsamples)
+    s2 = stats(samplesA.stats(nsamples=nsamples),
+               samplesB.stats(nsamples=nsamples),
+               samplesAB.stats(nsamples=nsamples))
+    assert_series_equal(s2.mean(), s.mean(), atol=s2.std().max())
 
 
 def test_tension_stats_incompatible_gaussian():
@@ -95,27 +98,29 @@ def test_tension_stats_incompatible_gaussian():
 
     nsamples = 10
     beta = 1
-    samples_stats = stats(samplesA, samplesB, samplesAB, nsamples, beta)
+    s = stats(samplesA, samplesB, samplesAB, nsamples, beta)
 
-    logR_std = samples_stats.logR.std()
-    logR_mean = samples_stats.logR.mean()
     logR_exact = logV - dmu_cov_dmu_AB/2 - slogdet(2*np.pi*(covA+covB))[1]/2
-    assert logR_mean == approx(logR_exact, abs=3*logR_std)
+    assert s.logR.mean() == approx(logR_exact, abs=3*s.logR.std())
 
-    logS_std = samples_stats.logS.std()
-    logS_mean = samples_stats.logS.mean()
     logS_exact = d / 2 - dmu_cov_dmu_AB / 2
-    assert logS_mean == approx(logS_exact, abs=3*logS_std)
+    assert s.logS.mean() == approx(logS_exact, abs=3*s.logS.std())
 
-    logI_std = samples_stats.logI.std()
-    logI_mean = samples_stats.logI.mean()
     logI_exact = logV - d / 2 - slogdet(2*np.pi*(covA+covB))[1] / 2
-    assert logI_mean == approx(logI_exact, abs=3*logI_std)
-
-    assert logS_mean == approx(logR_mean - logI_mean, abs=3*logS_std)
-
-    assert samples_stats.get_labels().tolist() == ([r'$\ln\mathcal{R}$',
-                                                    r'$\ln\mathcal{I}$',
-                                                    r'$\ln\mathcal{S}$',
-                                                    r'$d_\mathrm{G}$',
-                                                    r'$p$'])
+    assert s.logI.mean() == approx(logI_exact, abs=3*s.logI.std())
+
+    assert s.logS.mean() == approx(s.logR.mean() - s.logI.mean(),
+                                   abs=3*s.logS.std())
+
+    assert s.get_labels().tolist() == ([r'$\ln\mathcal{R}$',
+                                        r'$\ln\mathcal{I}$',
+                                        r'$\ln\mathcal{S}$',
+                                        r'$d_\mathrm{G}$',
+                                        r'$p$'])
+
+    with raises(ValueError):
+        stats(samplesA.stats(nsamples=5), samplesB, samplesAB, nsamples)
+    s2 = stats(samplesA.stats(nsamples=nsamples),
+               samplesB.stats(nsamples=nsamples),
+               samplesAB.stats(nsamples=nsamples))
+    assert_series_equal(s2.mean(), s.mean(), atol=s2.std().max())