update to newer version of seaborn

Author: SilasK

statsplot/dimred.py

@@ -133,14 +133,14 @@ def __init__(
         self, data, method=PCA, transformation=None, n_components=None, **kargs
     ):
 
-        if n_components is None:
-            n_components = data.shape[0]
-
         if data.shape[0] > data.shape[1]:
-            print(
+            warnings.warn(
                 "you don't need to reduce dimensionality or your dataset is transposed."
             )
 
+        if n_components is None:
+            n_components = min(data.shape)
+
         self.decomposition = method(n_components=n_components, **kargs)
 
         self.rawdata = data
@@ -154,7 +154,7 @@ def __init__(
 
         else:
 
-            self.data_ = data.applymap(transformation)
+            self.data_ = data.map(transformation)
 
         Xt = self.decomposition.fit_transform(self.data_)
 

statsplot/plot.py

@@ -189,8 +189,10 @@ def vulcanoplot(
         ax.annotate(g1, (ax_lim * 0.9, 0), ha="right")
         ax.annotate(g2, (-ax_lim * 0.9, 0), ha="left")
 
+
 # TODO: handle unaligned input.
 
+
 def statsplot(
     variable,
     test_variable,
@@ -261,7 +263,7 @@ def statsplot(
     if show_dots:
         legend = ax.get_legend_handles_labels()
 
-        sns.swarmplot(**params, color="k", dodge=True, **swarm_params)
+        sns.swarmplot(**params, palette="dark:k", dodge=True, **swarm_params)
 
         # add old variable, as dots have unified colors
         if grouping_variable is not None:

statsplot/siglabels.py

@@ -75,7 +75,7 @@ def __plot_sig_labels_hue(
     # start with y0
     y = y0
 
-    for idx, text in P_values.iteritems():
+    for idx, text in P_values.items():
 
         def calculate_hue_offset(group, order):
             return (order.index(group) - len(order) * 0.5 + 0.5) / len(order) * width
@@ -112,7 +112,7 @@ def ___plot_sig_labels_xaxis(
     P_values = P_values.apply(format_p_value, use_stars=use_stars)
 
     y = y0
-    for idx, text in P_values.iteritems():
+    for idx, text in P_values.items():
 
         def calculate_x_offset(group, order):
             return order.index(group)
@@ -135,7 +135,6 @@ def plot_all_sig_labels(
     ax=None,
     **kws,
 ):
-
     """"""
 
     # define y0 and deltay

statsplot/stats.py

@@ -77,6 +77,7 @@ def __stats_test_all_on_once(values1, values2, test, **test_kws):
     res = test(values1, values2, **test_kws)
     ResultsDB["Statistic"] = res.statistic
     ResultsDB["Pvalue"] = res.pvalue
+
     return ResultsDB
 
 
@@ -109,9 +110,9 @@ def two_group_test(
     test_kws=None,
     correct_for_multiple_testing=True,
 ):
-
     """test: a parwise statistical test found in scipy e.g ['mannwhitneyu','ttest_ind']
-    or a function wich takes two argumens. Additional keyword arguments can be specified by test_kws"""
+    or a function wich takes two argumens. Additional keyword arguments can be specified by test_kws
+    """
 
     # Define test
     if test_kws is None:
@@ -189,6 +190,9 @@ def two_group_test(
         Pairwise_comp = Test(values1, values2)
 
         Pairwise_comp["median_diff"] = values2.median() - values1.median()
+        Pairwise_comp["mean_diff"] = values2.mean() - values1.mean()
+        Pairwise_comp["Median1"] = values1.median()
+        Pairwise_comp["Median2"] = values2.median()
 
         if min_value >= 0:
             Pairwise_comp["log2FC"] = np.log2(values2.mean() + log_delta) - np.log2(
@@ -230,7 +234,6 @@ def calculate_stats(
     test="ttest_ind",
     **test_kws,
 ):
-
     """Calculate pairewise statistical tests optioonally grouped by a grouping variable"""
 
     kws = dict(

statsplot/statstable.py

@@ -128,14 +128,14 @@ def subset(self, index=None, columns=None):
 
     def groupby(self, groupby, axis=0):
         if axis == 0:
-            G = self.obs.groupby(groupby, axis=0)
+            G = self.obs.groupby(groupby)
 
             for group in G.indices:
                 yield (group, self.subset(index=self.obs_names[G.indices[group]]))
 
         elif axis == 1:
             # Group by on axis 0. var indexes contain data.columns
-            G = self.var.groupby(groupby, axis=0)
+            G = self.var.groupby(groupby)
             for group in G.indices:
                 yield group, self.subset(columns=self.var_names[G.indices[group]])
         else:
@@ -397,8 +397,7 @@ def __get_comparisons(self, subset=None):
 
             return list(subset)
 
-
-# TODO: Hide output axes labesls
+    # TODO: Hide output axes labesls
     def vulcanoplot(
         self,
         comparisons=None,

statsplot/transformations.py

@@ -1,12 +1,9 @@
-
-
-
 from numpy import log
 import pandas as pd
 import numpy as np
 
 
-# copied from scikit-bio 
+# copied from scikit-bio
 # because I cannot install it
 def closure(mat):
     """
@@ -93,24 +90,25 @@ def multiplicative_replacement(mat, delta=None):
            [ 0.0625,  0.4375,  0.4375,  0.0625]])
     """
     mat = closure(mat)
-    z_mat = (mat == 0)
+    z_mat = mat == 0
 
     num_feats = mat.shape[-1]
     tot = z_mat.sum(axis=-1, keepdims=True)
 
     if delta is None:
-        delta = (1. / num_feats)**2
+        delta = (1.0 / num_feats) ** 2
 
     zcnts = 1 - tot * delta
     if np.any(zcnts) < 0:
-        raise ValueError('The multiplicative replacement created negative '
-                         'proportions. Consider using a smaller `delta`.')
+        raise ValueError(
+            "The multiplicative replacement created negative "
+            "proportions. Consider using a smaller `delta`."
+        )
     mat = np.where(z_mat, delta, zcnts * mat)
     return mat.squeeze()
 
 
-
-def clr(data: pd.DataFrame, log=log,features="all"):
+def clr(data: pd.DataFrame, log=log, features="all"):
     """
     Centered log ratio (CLR) with multiplicative replacement implemented in scikit-bio
     """
@@ -128,6 +126,8 @@ def clr(data: pd.DataFrame, log=log,features="all"):
     # Fill in zeros with multiplicative replacement
     matrix = multiplicative_replacement(matrix)
 
+    matrix = pd.DataFrame(matrix, index=d.index, columns=d.columns)
+
     # CLR
     matrix = log(matrix)
 
@@ -137,7 +137,7 @@ def clr(data: pd.DataFrame, log=log,features="all"):
         mean = matrix.mean(1)
 
     elif features.lower() == "nz":
-            
+
         mean = matrix[matrix != 0].mean(1)
     elif features.lower() == "iql":
         # use mean of features in interquartile range
@@ -147,9 +147,6 @@ def clr(data: pd.DataFrame, log=log,features="all"):
     else:
         raise Exception("features must be 'all', 'nz', or 'iql'")
 
-
     matrix = (matrix.T - mean).T
 
-    if type(data) == pd.DataFrame:
-
-        return pd.DataFrame(matrix, index=d.index, columns=d.columns)
+    return matrix