Merge b7750d6 into cf1093e

sci_analysis/__init__.py

@@ -16,7 +16,7 @@
 # from sci_analysis.analysis import analysis
 # from sci_analysis.graphs import graph
 # from sci_analysis.data import data
-__all__ = ["data", "analysis", "graph"]
+__all__ = ["data", "analysis", "graphs"]
 from .analysis import analyze, analyse
 # from .analysis.analysis import Comparison, NormTest, TTest, LinearRegression,\
 #     Correlation, Anova, Kruskal, EqualVariance, VectorStatistics, GroupStatistics

sci_analysis/analysis/__init__.py

@@ -248,16 +248,19 @@ def analyze(xdata, ydata=None, groups=None, labels=None, alpha=0.05, **kwargs):
             tested.append('Distribution')
 
             # Show the histogram and stats
-            if 'sample' in kwargs:
-                out_stats = VectorStatistics(_data, sample=kwargs['sample'], display=False)
-            else:
-                out_stats = VectorStatistics(_data, display=False)
+            out_stats = VectorStatistics(_data, sample=kwargs.get('sample', False), display=False)
             if 'distribution' in kwargs:
                 distro = kwargs['distribution']
-                distro_class = getattr(__import__('scipy.stats',
-                                                  globals(),
-                                                  locals(),
-                                                  [distro], 0), distro)
+                distro_class = getattr(
+                    __import__(
+                        'scipy.stats',
+                        globals(),
+                        locals(),
+                        [distro],
+                        0,
+                    ),
+                    distro,
+                )
                 parms = distro_class.fit(xdata)
                 fit = KSTest(xdata, distribution=distro, parms=parms, alpha=alpha, display=False)
                 tested.append('KSTest')

sci_analysis/analysis/hypo_tests.py

@@ -237,7 +237,7 @@ def run(self):
         args = self._data.groups.values()
         if len(args) <= 1:
             raise NoDataError("At least one of the inputs is empty or non-numeric.")
-        u_value, p_value = mannwhitneyu(*args)
+        u_value, p_value = mannwhitneyu(*args, alternative='less')
         self._results.update({self._statistic_name: u_value, 'p value': p_value * 2, 'alpha': self._alpha})
 
     @property

sci_analysis/analysis/stats.py

@@ -156,8 +156,8 @@ class GroupStatistics(Analysis):
     _num_of_groups = 'Number of Groups'
 
     def __init__(self, *args, **kwargs):
-        groups = kwargs['groups'] if 'groups' in kwargs else None
-        display = kwargs['display'] if 'display' in kwargs else True
+        groups = kwargs.get('groups', None)
+        display = kwargs.get('display', False)
         if is_dict(args[0]):
             _data, = args
         elif is_group(args,):

sci_analysis/graphs/vector.py

@@ -79,22 +79,18 @@ def __init__(self, data, **kwargs):
         :param _save_to: Save the graph to the specified path.
         :return: pass
         """
-        self._bins = kwargs['bins'] if 'bins' in kwargs else 20
-        self._distribution = kwargs['distribution'] if 'distribution' in kwargs else 'norm'
-        self._box_plot = kwargs['boxplot'] if 'boxplot' in kwargs else True
-        self._cdf = kwargs['cdf'] if 'cdf' in kwargs else False
-        self._fit = kwargs['fit'] if 'fit' in kwargs else False
-        self._mean = kwargs['mean'] if 'mean' in kwargs else None
-        self._std = kwargs['std_dev'] if 'std_dev' in kwargs else None
-        self._sample = kwargs['sample'] if 'sample' in kwargs else True
-        self._title = kwargs['title'] if 'title' in kwargs else 'Distribution'
-        self._save_to = kwargs['save_to'] if 'save_to' in kwargs else None
-        yname = "Probability"
-        name = 'Data'
-        if 'name' in kwargs:
-            name = kwargs['name']
-        elif 'xname' in kwargs:
-            name = kwargs['xname']
+        self._bins = kwargs.get('bins', 20)
+        self._distribution = kwargs.get('distribution', 'norm')
+        self._box_plot = kwargs.get('boxplot', True)
+        self._cdf = kwargs.get('cdf', False)
+        self._fit = kwargs.get('fit', False)
+        self._mean = kwargs.get('mean')
+        self._std = kwargs.get('std_dev')
+        self._sample = kwargs.get('sample', False)
+        self._title = kwargs.get('title', 'Distribution')
+        self._save_to = kwargs.get('save_to')
+        yname = kwargs.get('yname', 'Probability')
+        name = kwargs.get('name') or kwargs.get('xname') or 'Data'
 
         super(GraphHisto, self).__init__(data, xname=name, yname=yname)
 
@@ -110,10 +106,16 @@ def fit_distro(self):
             Third value - The cdf y-axis points
 
         """
-        distro_class = getattr(__import__('scipy.stats',
-                                          globals(),
-                                          locals(),
-                                          [self._distribution], 0), self._distribution)
+        distro_class = getattr(
+            __import__(
+                'scipy.stats',
+                globals(),
+                locals(),
+                [self._distribution],
+                0,
+            ),
+            self._distribution
+        )
         parms = distro_class.fit(self._data.data)
         distro = linspace(distro_class.ppf(0.001, *parms), distro_class.ppf(0.999, *parms), 100)
         distro_pdf = distro_class.pdf(distro, *parms)
@@ -167,7 +169,7 @@ def draw(self):
 
         # Set the title
         title = self._title
-        if self._mean is not None and self._std is not None:
+        if self._mean and self._std:
             if self._sample:
                 title = r"{}{}$\bar x = {:.4f},  s = {:.4f}$".format(title, "\n", self._mean, self._std)
             else:

sci_analysis/test/test_analyze.py

@@ -261,13 +261,13 @@ def test_124_distribution_label(self):
                                  save_to='{}test_analyze_124'.format(self.save_path)),
                          ['Distribution', 'NormTest'])
 
-    def test_125_distribution_population(self):
-        """Perform a distribution analysis with population set"""
+    def test_125_distribution_sample(self):
+        """Perform a distribution analysis with sample set"""
         np.random.seed(self._seed)
         input_array = st.norm.rvs(size=200)
         self.assertEqual(analyze(input_array,
-                                 sample=False,
-                                 title='Population Stats',
+                                 sample=True,
+                                 title='Sample Stats',
                                  debug=True,
                                  save_to='{}test_analyze_125'.format(self.save_path)),
                          ['Distribution', 'NormTest'])

sci_analysis/test/test_graph_histo.py

@@ -157,7 +157,7 @@ def test_114_mean_std_and_sample(self):
         self.assertTrue(GraphHisto(input_array,
                                    mean=mean,
                                    std_dev=std,
-                                   sample=False,
+                                   sample=True,
                                    save_to='{}test_histo_114'.format(self.save_path)))
 
     def test_115_distribution(self):

sci_analysis/test/test_mannwhitney.py

@@ -23,15 +23,15 @@ def test_MannWhitney_matched(self):
 -------------------
 
 alpha   =  0.0500
-u value =  4976.0000
-p value =  0.9542
+u value =  5024.0000
+p value =  1.0477
 
 H0: Locations are matched
 """
         self.assertGreater(exp.p_value, alpha, "FAIL: MannWhitney Type I error")
-        self.assertAlmostEqual(exp.statistic, 4976.0, delta=0.0001, msg="FAIL: MannWhitney statistic incorrect")
-        self.assertAlmostEqual(exp.u_value, 4976.0, delta=0.0001, msg="FAIL: MannWhitney u_value incorrect")
-        self.assertAlmostEqual(exp.p_value, 0.9542, delta=0.0001, msg="FAIL: MannWhitney p_value incorrect")
+        self.assertAlmostEqual(exp.statistic, 5024.0, delta=0.0001, msg="FAIL: MannWhitney statistic incorrect")
+        self.assertAlmostEqual(exp.u_value, 5024.0, delta=0.0001, msg="FAIL: MannWhitney u_value incorrect")
+        self.assertAlmostEqual(exp.p_value, 1.0477, delta=0.0001, msg="FAIL: MannWhitney p_value incorrect")
         self.assertEqual(str(exp), output)
 
     def test_MannWhitney_unmatched(self):
@@ -75,8 +75,8 @@ def test_MannWhitney_matched_just_above_min_size(self):
 -------------------
 
 alpha   =  0.0500
-u value =  219.0000
-p value =  0.9799
+u value =  222.0000
+p value =  1.0401
 
 H0: Locations are matched
 """
@@ -144,15 +144,15 @@ def test_MannWhitney_vector_input(self):
 -------------------
 
 alpha   =  0.0500
-u value =  4976.0000
-p value =  0.9542
+u value =  5024.0000
+p value =  1.0477
 
 H0: Locations are matched
 """
         self.assertGreater(exp.p_value, alpha, "FAIL: MannWhitney Type I error")
-        self.assertAlmostEqual(exp.statistic, 4976.0, delta=0.0001, msg="FAIL: MannWhitney statistic incorrect")
-        self.assertAlmostEqual(exp.u_value, 4976.0, delta=0.0001, msg="FAIL: MannWhitney u_value incorrect")
-        self.assertAlmostEqual(exp.p_value, 0.9542, delta=0.0001, msg="FAIL: MannWhitney p_value incorrect")
+        self.assertAlmostEqual(exp.statistic, 5024.0, delta=0.0001, msg="FAIL: MannWhitney statistic incorrect")
+        self.assertAlmostEqual(exp.u_value, 5024.0, delta=0.0001, msg="FAIL: MannWhitney u_value incorrect")
+        self.assertAlmostEqual(exp.p_value, 1.0477, delta=0.0001, msg="FAIL: MannWhitney p_value incorrect")
         self.assertEqual(str(exp), output)
 
     def test_MannWhitney_missing_second_arg(self):