adapting code for empirical graphs

CSHVienna · Apr 26, 2023 · 42d2a80 · 42d2a80
1 parent 26fa88b
commit 42d2a80
Show file tree

Hide file tree

Showing 3 changed files with 18 additions and 6 deletions.
diff --git a/netin/__init__.py b/netin/__init__.py
@@ -4,7 +4,7 @@
 in social networks.
 """
 
-__version__ = '1.0.5.9'
+__version__ = '1.0.5.9.1'
 
 from netin import generators
 from netin.generators import *

diff --git a/netin/generators/graph.py b/netin/generators/graph.py
@@ -754,15 +754,23 @@ def get_node_metadata_as_dataframe(self, include_graph_metadata: bool = False, n
         -------
         pd.DataFrame
             dataframe with the metadata of the nodes
+
+        Notes
+        -----
+        Column `class_label` is a binary column indicating whether the node belongs to the minority class.
         """
-        cols = ['node', 'class_label']
+        cols = ['node', 'class_label', 'real_label', 'source']
+
         obj = {'node': self.node_list,
-               'class_label': [self.get_class_label(n) for n in self.node_list]}
+               'class_label': [const.MAJORITY_LABEL if self.get_class_label(n) == self.get_majority_label() else const.MINORITY_LABEL for n in self.node_list],
+               'real_label': [self.get_class_label(n) for n in self.node_list],
+               'source': 'model' if 'empirical' not in self.graph else 'data'}
 
         # include graph metadata
         if include_graph_metadata:
             n = self.number_of_nodes()
-            newcols = [c for c in self.graph.keys() if c not in ['class_attribute', 'class_values', 'class_labels']]
+            newcols = [c for c in self.graph.keys() if c not in ['class_attribute', 'class_values',
+                                                                 'class_labels']]
             obj.update({c: self.graph[c] for c in newcols})
             cols.extend(newcols)
 

diff --git a/netin/viz/handlers.py b/netin/viz/handlers.py
@@ -11,6 +11,7 @@
 import pandas as pd
 import seaborn as sns
 from matplotlib import rc
+from collections import Counter
 
 from netin.generators.graph import Graph
 from netin.stats.distributions import fit_power_law
@@ -101,7 +102,9 @@ class label e.g., M or m for majority and minority
     -------
     color: str
         color for the class label
+
     """
+
     if ylabel in MINORITY_CURVE and class_label is None:
         return COLOR_MINORITY
 
@@ -458,7 +461,8 @@ def plot_distribution(data: Union[pd.DataFrame, List[pd.DataFrame]],
             total = df[_col_name].sum() if common_norm else group[_col_name].sum()
             xs, ys = get_x_y_from_df_fnc(group, _col_name, total)
             plot = ax.scatter if scatter else ax.plot
-            plot(xs, ys, label=class_label, color=_get_class_label_color(class_label, xy_fnc_name), **kwargs)
+            plot(xs, ys, label=class_label, color=_get_class_label_color(class_label=class_label,
+                                                                         ylabel=xy_fnc_name), **kwargs)
 
             if hline_fnc:
                 hline_fnc(ax.axhline, group)
@@ -789,7 +793,7 @@ def plot_powerlaw_fit(data: Union[pd.DataFrame, List[pd.DataFrame]], col_name: U
             discrete = group_nonzero[_col_name].dtype == np.int64
             fit = fit_power_law(group_nonzero.loc[:, _col_name].values, discrete=discrete, verbose=verbose)
 
-            color = _get_class_label_color(class_label)
+            color = _get_class_label_color(class_label=class_label)
 
             efnc = fit.plot_ccdf if kind == "ccdf" else fit.plot_cdf if kind == 'cdf' else fit.plot_pdf
             fnc = fit.power_law.plot_ccdf if kind == "ccdf" else fit.power_law.plot_cdf if kind == 'cdf' \