Add count anonymity method

anonypy/anonypy.py

@@ -23,6 +23,24 @@ def anonymize_l_diversity(self, k, l):
     def anonymize_t_closeness(self, k, p):
         return self.__anonymize(k, p=p)
 
+    def __count_anonymity(self, k, l=0, p=0.0):
+        partitions = self.modrian.partition(k, l, p)
+        return count_anonymity(
+            self.modrian.df,
+            partitions,
+            self.modrian.feature_columns,
+            self.modrian.sensitive_column,
+        )
+
+    def count_k_anonymity(self, k):
+        return self.__count_anonymity(k)
+
+    def count_l_diversity(self, k, l):
+        return self.__count_anonymity(k, l=l)
+
+    def count_t_closeness(self, k, p):
+        return self.__count_anonymity(k, p=p)
+
 
 def agg_categorical_column(series):
     # this is workaround for dtype bug of series
@@ -50,8 +68,6 @@ def anonymize(df, partitions, feature_columns, sensitive_column, max_partitions=
             aggregations[column] = agg_numerical_column
     rows = []
     for i, partition in enumerate(partitions):
-        if i % 100 == 1:
-            print("Finished {} partitions...".format(i))
         if max_partitions is not None and i > max_partitions:
             break
         grouped_columns = df.loc[partition].agg(aggregations, squeeze=False)
@@ -70,3 +86,23 @@ def anonymize(df, partitions, feature_columns, sensitive_column, max_partitions=
             )
             rows.append(values.copy())
     return rows
+
+
+def count_anonymity(
+    df, partitions, feature_columns, sensitive_column, max_partitions=None
+):
+    aggregations = {}
+    for column in feature_columns:
+        if df[column].dtype.name == "category":
+            aggregations[column] = agg_categorical_column
+        else:
+            aggregations[column] = agg_numerical_column
+    aggregations[sensitive_column] = "count"
+    rows = []
+    for i, partition in enumerate(partitions):
+        if max_partitions is not None and i > max_partitions:
+            break
+        grouped_columns = df.loc[partition].agg(aggregations, squeeze=False)
+        values = grouped_columns.iloc[0].to_dict()
+        rows.append(values.copy())
+    return rows

setup.py

@@ -8,14 +8,15 @@
 
 setup(
     name="anonypy",
-    version="0.1.1",
+    version="0.1.4",
     packages=find_packages(),
     author="glassonion1",
     author_email="[email protected]",
+    url="https://github.com/glassonion1/anonypy",
     description="Anonymization library for python",
     long_description=long_description,
     long_description_content_type="text/markdown",
-    keywords="k-anonymity l-diversity t-closeness",
+    keywords="k-anonymity l-diversity t-closeness mondrian",
     classifiers=[
         "Programming Language :: Python :: 3.8",
         "Programming Language :: Python :: 3.9",

tests/preserver_test.py

@@ -34,6 +34,23 @@ def test_k_anonymity():
     print(dfn)
 
 
+def test_count_k_anonymity():
+    df = pd.DataFrame(data=data, columns=columns)
+    print(df)
+
+    for name in categorical:
+        df[name] = df[name].astype("category")
+
+    feature_columns = ["col1", "col2", "col3"]
+    sensitive_column = "col4"
+
+    p = anonypy.Preserver(df, feature_columns, sensitive_column)
+    rows = p.count_k_anonymity(k=2)
+
+    dfn = pd.DataFrame(rows)
+    print(dfn)
+
+
 def test_l_diversity():
     df = pd.DataFrame(data=data, columns=columns)