tfdv wrapper file

cmflib/cmf_analysis.py

@@ -0,0 +1,30 @@
+import tensorflow_data_validation as tfdv
+import pandas as pd
+class CmfAnalysis():
+    def __init__(self):
+        pass
+
+    def create_statistics(self,data_path,feature_stats_path):
+        statistics=tfdv.generate_statistics_from_csv(data_path)
+        tfdv.write_stats_text(statistics, feature_stats_path)
+
+    def visualize_statistics(self,statistics):
+        stats=tfdv.load_statistics(statistics)
+        tfdv.visualize_statistics(stats)
+
+    def create_schema(self,data_path,schema_path):
+        schema=tfdv.infer_schema(statistics=tfdv.generate_statistics_from_csv(data_path,delimiter=";"))
+        tfdv.write_schema_text(schema,schema_path)
+
+    def validate_data(self,train_stats_path,eval_path,schema_path):
+        new_csv_stats = tfdv.load_statistics(train_stats_path)
+        eval_stats=tfdv.load_statistics(eval_path)
+        schema=tfdv.load_schema_text(schema_path)
+        anomalies = tfdv.validate_statistics(statistics=eval_stats, schema=schema)
+        return anomalies
+
+    def display_anomalies(self,anomalies):
+        tfdv.display_anomalies(anomalies)
+
+if __name__=="__main__":
+    validator=CmfAnalysis()