Update interface.py

clust/ML/clustering/interface.py

@@ -85,4 +85,64 @@ def clusteringByMethod(data, parameter, model_path):
     return data_series, result, plt1
 
 
+def clusteringByCDM(data, parameter, model_path):
+    """ make clustering result of multiple dataset 
+
+    Retrieves rows pertaining to the given keys from the Table instance
+    represented by table_handle.  String keys will be UTF-8 encoded.
+
+    Args:
+        data (dataFrame): list of multiple dataframe inputs to be clustered (each column : individual data, index : time sequence data)
+        model(int): clust model name to be applied modelList = ['som']
+        x(int): x length
+        y(int): y length
+
+    Returns:
+        Dictionary: result (A dict mapping keys to the corresponding clustering result number)
+    
+    **Return Result Example**::
+
+        result = { b'ICW0W2000011': '5',
+                   b'ICW0W2000013': '4',
+                   b'ICW0W2000014': '6'... }
+    """
+
+    result = None
+    param = parameter['param']
+    model_name = parameter['method']
+    data_series = ''
+    if (len(data.columns) > 0):
+        # Train/Test 클래스 생성
+        # 모델 저장/로드 경로
+        if model_name == "som":
+            clust_train = SomTrain()
+            clust_train.set_param(param)
+            clust_test = SomTest()
+        elif model_name == "kmeans":
+            clust_train = KMeansTrain()
+            clust_train.set_param(param)
+            clust_test = KMeansTest()
+
+        # 1. 데이터 준비
+        data_series = DF_to_series(data)
+
+
+        # 2. Train
+        clust_train.set_param(param)
+        clust_train.train(data_series)
+
+        # 3. model save
+        save_pickle_model(clust_train.model, model_path)
+
+        # 4. model load
+        model = load_pickle_model(model_path)
+
+        # 5. predict
+        clust_test.set_model(model)
+        result = clust_test.predict(data_series)
+
+    else:
+        result = None
+
+    return data_series, result