allow select columns

gapp.py

@@ -23,20 +23,22 @@ logger = logging.getLogger(__name__)
 logger.setLevel(logging.DEBUG)
 
 
-def get_data(temp_file, state: dict):
+def get_data(temp_file, case_col, activity_col, timestamp_col, state: dict):
     # print(f"temp_file: {temp_file}")
     if isinstance(temp_file, str):
     #   df = pd.read_csv(StringIO(temp_file), parse_dates=[ "Start", "Finish"])
-        df = pd.read_csv(temp_file, sep=';', dtype={'case_id': str}, parse_dates = ['timestamp'])
-        df.loc[:, 'timestamp'] = pd.to_datetime(df['timestamp']) # format='%Y-%m-%d %H:%M:%S'
+        df = pd.read_csv(temp_file, sep=';|,')
     else:
         # df = pd.read_csv(temp_file.name, ) # parse_dates=[ "Start", "Finish"]
-        df = pd.read_csv(temp_file.name, sep=';', dtype={'case_id': str}, parse_dates = ['timestamp'])
-        df.loc[:, 'timestamp'] = pd.to_datetime(df['timestamp'])
+        df = pd.read_csv(temp_file.name, sep=';|,')
     # logger.debug(df.head())
     # logger.debug(df.dtypes)
     state['df'] = df
-    return df, state
+    return df, \
+        gr.Dropdown( choices=list(df.columns), multiselect=False, label="Case", info="選擇 Case ID"),  \
+        gr.Dropdown( choices=list(df.columns), multiselect=False, label="Activity", info="選擇 Activity ID"),  \
+        gr.Dropdown( choices=list(df.columns), multiselect=False, label="Timestamp", info="選擇 Timestamp"), \
+        state
 
 
 def get_stats(state: dict):
@@ -101,6 +103,20 @@ def get_process_map_activities_connections( activity_rank: int = 0, connection_r
     return chart, state
 
 
+def etl( case_col, activity_col, timestamp_col, state: dict):
+    """
+    Argument
+    Return
+    """
+    df = state['df'].copy()
+    df.loc[:, case_col] = df[case_col].astype(str)
+    df.loc[:, activity_col] = df[activity_col].astype(str)
+    df.loc[:, timestamp_col] = pd.to_datetime(df[timestamp_col]) # format='%Y-%m-%d %H:%M:%S'
+    df.rename(columns={case_col: 'case_id', activity_col: 'activity', timestamp_col: 'timestamp'}, inplace=True)
+    state['df'] = df
+    return df, state
+
+
 ## --- block --- ##
 css = """
 h1 {
@@ -114,11 +130,19 @@ with demo:
     state = gr.State(value={})
     with gr.Row():
         upl_btn = gr.UploadButton(label="Upload", file_types = ['.csv'], file_count = "single")
-    # with gr.Row('Data Preview'):
     with gr.Accordion('Data Preview'): 
         df = gr.Dataframe()
-    upl_btn.upload( fn=get_data, inputs = [upl_btn, state], outputs=[df, state])
-    
+
+    with gr.Row():
+        case_col = gr.Dropdown( multiselect=False, label="Case", info="選擇 Case ID")
+        activity_col = gr.Dropdown( multiselect=False, label="Activity", info="選擇 Activity ID")
+        timestamp_col = gr.Dropdown( multiselect=False, label="Timestamp", info="選擇 Timestamp")
+
+    upl_btn.upload( fn=get_data, inputs = [upl_btn, case_col, activity_col, timestamp_col, state], outputs=[df, case_col, activity_col, timestamp_col, state])
+
+    column_btn = gr.Button("Select Columns")
+    column_btn.click( fn=etl, inputs = [ case_col, activity_col, timestamp_col, state], outputs=[df, state])
+
     with gr.Row():
         with gr.Tab('Data Explorer'):
             # outputs.append(gr.Dataframe( label="Event logs"))
@@ -129,6 +153,7 @@ with demo:
                 chart2 = gr.BarPlot( label="Case Lead Time Stats")
                 chart3 = gr.BarPlot( label="Case Average Activity Time Stats")
             de_btn.click( fn=get_stats, inputs = [state], outputs=[ summary, chart1, chart2, chart3, state])
+
         with gr.Tab('Variant Explorer'):
             ve_btn = gr.Button("Get Variants")
             top_k_variant_selector = gr.Slider(0, 10, value=1, step=1, label="Top-K", info="選擇 Variant 數量（0: 全選）")