add table display

Author: kinkunchan

.DS_Store

@@ -1,4 +1,4 @@
-from ..exceptions import UnimplementedException, MethodKeywordUnAllowedException
+from ..exceptions import *
 import networkx as nx
 import numpy as np
 import sklearn
@@ -33,7 +33,7 @@ def get_centrality(graph, centrality='degree', **kwargs):
     """
 
     if not isinstance(graph, nx.classes.graph.Graph):
-        raise InputFormatError("Please input graph as NetworkX.graph object")
+        raise InputFormatErrorException("Please input graph as NetworkX.graph object")
 
     if centrality == 'degree':
         dict_centrality = dict(graph.degree())
@@ -67,7 +67,7 @@ def centrality_correlation(graph, dict_embeddings, centrality='degree', similari
     """
 
     if not isinstance(graph, nx.classes.graph.Graph):
-        raise InputFormatError("Please input graph as NetworkX.graph object")
+        raise InputFormatErrorException("Please input graph as NetworkX.graph object")
 
     if centrality == 'degree':
         dict_centrality = dict(graph.degree())

Example_new.ipynb

@@ -1,4 +1,4 @@
-from ..exceptions import UnimplementedException, MethodKeywordUnAllowedException
+from ..exceptions import *
 
 import networkx as nx
 import numpy as np

semb/.DS_Store

@@ -1,4 +1,4 @@
-from ..exceptions import UnimplementedException, MethodKeywordUnAllowedException
+from ..exceptions import *
 
 import networkx as nx
 import numpy as np
@@ -19,12 +19,8 @@ def purity_score(y_true, y_pred):
 def kmeans_best_result(X, y, n_clusters):
     list_purity = list()
     list_nmi = list()
-    
     kmeans = KMeans(n_clusters=n_clusters, n_init=min(len(y), 1000), init='k-means++').fit(X)
-    
-    list_purity += [purity_score(y, kmeans.labels_)]
-    list_nmi += [normalized_mutual_info_score(y, kmeans.labels_)]
-    return {'purity': list_purity, 'nmi': list_nmi}
+    return {'purity': purity_score(y, kmeans.labels_), 'nmi': normalized_mutual_info_score(y, kmeans.labels_)}
 
 
 def perform_clustering(dict_embeddings, dict_labels, **kwargs):

semb/evaluations/centrality_correlation.py

@@ -1,5 +1,6 @@
-from ..exceptions import UnimplementedException, MethodKeywordUnAllowedException
+from ..exceptions import *
 import networkx as nx
+import pandas as pd
 
 def get_label(input_dir, delimeter = ' ' ,**kwargs):
     """
@@ -33,3 +34,56 @@ def get_label(input_dir, delimeter = ' ' ,**kwargs):
     for key, val in dict_counter.items():
         print(">>> Label", key, 'appears', val, 'times')
     return dict_labels
+
+def concatenate_result_pd(list_results):
+    """
+    Concatenate the results from the clustering / classifcation test
+    Arguments:
+    list_results {list of tuples} --  [("method_name", dict_result)], where the dict_result is the returned dict
+                                      from the perform_clustering() and perform_classification() functions
+    
+    Return:
+    pd_results -- a pandas table showing the results
+
+    """
+    # Perform input checking on the list_results
+    if len(list_results) == 0:
+        raise InputFormatErrorException("Input length 0!")
+
+    for cur_item in list_results:
+        if (len(cur_item) != 2):
+            raise InputFormatErrorException("Please input the results as list of tuples, i.e. [(\"method_name\", dict_result)]")
+
+        if (not isinstance(cur_item[0], str)) or (not isinstance(cur_item[1], dict)):
+            raise InputFormatErrorException("Please input the results as list of tuples, i.e. [(\"method_name\", dict_result)]")
+
+        if "overall" not in cur_item[1]:
+            raise InputFormatErrorException("Invalid input. Please make sure that the input result is generated from perform_classification() or perform_clustering()")
+
+    
+    pd_results = pd.DataFrame()
+    pd_results['methods'] = [i[0] for i in list_results]
+
+
+    
+    # Peform checking on whether classifcation or clustering is tested
+    if 'accuracy' in list_results[0][1]['overall']:
+        # Classification
+        for cur_item in list_results:
+            if 'accuracy' not in cur_item[1]['overall']:
+                raise InputFormatErrorException("Invalid input. Please make sure that the input result is generated from perform_classification()")
+                
+        for metric in ['accuracy', 'f1_macro', 'f1_micro', 'auc_micro', 'auc_macro']:
+            for value in ['mean', 'std']:
+                pd_results[metric + '_' + value] = [i[1]['overall'][metric][value] for i in list_results]
+    else:
+        # Clustering
+        for cur_item in list_results:
+            if 'purity' not in cur_item[1]['overall']:
+                raise InputFormatErrorException("Invalid input. Please make sure that the input result is generated from perform_clustering()")
+        
+        for metric in ['purity', 'nmi']:
+            pd_results[metric] = [i[1]['overall'][metric] for i in list_results]
+        
+    return pd_results
+

semb/evaluations/classification.py

@@ -10,5 +10,5 @@ class MethodNotExistException(Exception):
 class MethodKeywordUnAllowedException(Exception):
     pass
 
-class InputFormatError(Exception):
+class InputFormatErrorException(Exception):
     pass