Update template_gradio_interface.py

template_gradio_interface.py

@@ -14,33 +14,7 @@ import yaml
 # import time
 # import cv2
 # from prediction_coatings import predict
-
-
-def predict(
-    model_path,
-    data,
-    encoder,
-    minmax_scaler_inputs,
-    minmax_scaler_targets,
-    categorical_columns,
-    numerical_columns,
-    target_columns,
-    explainer,
-    use_in_gradio=False,
-):
-    # model = tf.keras.models.load_model(model_path)
-    # if use_in_gradio:
-    #     data = encode_categorical(data, categorical_columns, encoder=one_hot_scaler, fit=False)
-    #     data = scale_numerical(data, numerical_columns, scaler=minmax_scaler_inputs, fit=False)
-    # return model.predict(data), data.columns, explainer.shap_values(data[-10:])
-    return 34, 10, 45, 10, None
-
-
-def unpickle_file(path):
-    with open(path, "rb") as file:
-        unpickler = pickle.Unpickler(file)
-        unpickled_file = unpickler.load()
-    return unpickled_file
+from utils import predict, unpickle_file, scale_numertical, encode_categorical
 
 
 def call_predict(inference_dict, cols_order):
@@ -48,11 +22,6 @@ def call_predict(inference_dict, cols_order):
     Encapsulates the predict function from utils to pass the config, and to put the data in the right format
     """
 
-    categorical_columns = ""
-    numerical_columns = ""
-    target_columns = ""
-
-
     scaler_inputs = unpickle_file(inference_dict["inference"]["scaler_inputs_path"])
     scaler_targets = unpickle_file(inference_dict["inference"]["scaler_targets_path"])
     encoder = unpickle_file(inference_dict["inference"]["encoder_path"])
@@ -67,6 +36,8 @@ def call_predict(inference_dict, cols_order):
         df = pd.DataFrame([x_list], columns=cols_order)
         print(df.shape)
 
+        df_preprocessed = scale_numertical(df, numerical_columns, scaler=scaler_inputs, fit=False)
+        df_preprocessed = encode_categorical(df_preprocessed, categorical_columns, encoder=encoder, fit=False)
 
         y_pred, _, shap_values = predict(inference_dict["inference"]["model_path"], df_preprocessed, explainer)