新增了意见框、物理含义映射、index轴等功能

app.py

@@ -226,7 +226,7 @@ def process_salesforce_data(dataset: str, config: str, split: str, page: List[st
         tot_samples, tot_targets = max_page, len(df['target'][0]) if isinstance(df['target'][0], np.ndarray) and df['target'][0].dtype == 'O' else 1
         if 'all' in sub_targets:
             sub_targets = [i for i in range(tot_targets)]
-        df = clean_up_df(df, sub_targets)
+        df = clean_up_df(df, sub_targets, SUBTARGET_MEANING_MAP[config])
         row = df.iloc[0]
         id_list.append(row['item_id'])
         # 将单行的DataFrame展开为新的DataFrame
@@ -276,11 +276,16 @@ with gr.Blocks() as demo:
     #                   "statistics_textbox": statistics_textbox,
     #                   "user_input_box": user_input_box,
     #                   "plot": plot})
+    with gr.Row():
+        with gr.Column(scale=1):
+            choose_retain = gr.Dropdown(["delete", "retain", "modify"], label="Choose to retain or delete or modify", interactive=True)
+        with gr.Column(scale=2):
+            choose_retain_reason_box = gr.Textbox(label="Reason", placeholder="Enter your reason", interactive=True)
     score_slider = gr.Slider(1, 5, 1, step=0.5, label="Score for answer", interactive=True)
     with gr.Row():
         with gr.Column(scale=2):
             user_submit_button = gr.Button("submit", interactive=True)
-            user_name_box = gr.Textbox(label="user_name", placeholder="Enter your name first", interactive=True)
+            user_name_box = gr.Textbox(label="user_name", placeholder="Enter your name firstly", interactive=True)
         with gr.Column(scale=1):
             submit_info_box = gr.Textbox(label="submit_info", interactive=False)
     with gr.Row():
@@ -401,7 +406,7 @@ with gr.Blocks() as demo:
     cp_split.change(show_dataset_at_config_and_split, inputs=[cp_dataset, cp_config, cp_split], outputs=all_outputs)
     cp_goto_page.click(show_dataset_at_config_and_split_and_page, inputs=[cp_dataset, cp_config, cp_split, cp_page], outputs=all_outputs)
     cp_goto_next_page.click(show_dataset_at_config_and_split_and_next_page, inputs=[cp_dataset, cp_config, cp_split, cp_page], outputs=all_outputs)
-    user_submit_button.click(save_score, inputs=[user_name_box, cp_config, qusetion_id_box, score_slider], outputs=[submit_info_box])
+    user_submit_button.click(save_score, inputs=[user_name_box, cp_config, qusetion_id_box, score_slider, choose_retain, choose_retain_reason_box], outputs=[submit_info_box])
     # select_buttom.click(show_dataset_at_config_and_split_and_page, inputs=[cp_dataset, cp_config, cp_split, select_sample_box, select_subtarget_box], outputs=all_outputs)
 
 
@@ -412,3 +417,10 @@ if __name__ == "__main__":
     # import subprocess
     # subprocess.Popen(["python", "test_server.py"])
     uvicorn.run(app, host=host, port=7860)
+
+#// 对一下数据 --
+#// 部署到服务器上
+#// 测试一下功能 --
+#// 加一个选择文本框【删除、保留、修改】，加一个意见的文本框 --
+#// 横坐标增加一个代表index的轴 -
+#// 加一个物理含义的映射 -

comm_utils.py

@@ -9,14 +9,16 @@ def save_to_file(user_input):
     with open("user_input.txt", "w") as file:
         file.write(user_input)
 
-def save_score(user_id, subset_name, question_id, score):
+def save_score(user_id, subset_name, question_id, score, retain_option, retain_reason):
     if user_id.strip() == '':
         raise ValueError("User ID cannot be empty.")
     
     score_data = {
         "subset_name": subset_name,
         "question_id": question_id,
-        "score": score
+        "score": score,
+        "retain_option": retain_option,
+        "retain_reason": retain_reason
     }
 
     # 创建用户目录路径

config.py

@@ -15,4 +15,12 @@ COLUMN_ANSWER      = 'answer'
 COLUMN_DOMAIN      = 'domain'
 COLUMN_SOURCE      = 'source'
 COLUMN_LOCAL_OVERALL = 'local_overall'
-COLUMN_OPTIONS     = 'options'
+COLUMN_OPTIONS     = 'options'
+
+#subtarget meaning map
+SUBTARGET_MEANING_MAP = {'traffic_hourly'             : {0:'Road occupancy rates'},
+                         'alibaba_cluster_trace_2018' : {0:'CPU',
+                                                         1:'memory usage'},
+                         'era5_2018'                  : {**{k: 'Relative Humidity' for k in range(10, 17)},
+                                                        **{k: 'Temperature' for k in range(24, 31)}}}
+#TODO:不同索引下名称不能相同，否则生成df时会报错

utils.py

@@ -35,7 +35,7 @@ def ndarray_to_base64(ndarray):
     base64_str = base64.b64encode(buffer.getvalue()).decode('utf-8')
     return f"data:image/png;base64,{base64_str}"
 
-def flatten_ndarray_column(df, column_name, rows_to_include):
+def flatten_ndarray_column(df, column_name, rows_to_include, name_mapping_map:dict|None=None):
     """
     将嵌套的np.ndarray列展平为多列，并只保留指定的行。
     """
@@ -50,7 +50,10 @@ def flatten_ndarray_column(df, column_name, rows_to_include):
     selected_data = df[column_name].apply(select_and_flatten)
 
     for i, index in enumerate(rows_to_include):
-        df[f'{column_name}_{index}'] = selected_data.apply(lambda x: x[i])
+        if name_mapping_map is not None and index in name_mapping_map:
+            df[f'{column_name}_{name_mapping_map[index]}'] = selected_data.apply(lambda x: x[i])
+        else:
+            df[f'{column_name}_{index}'] = selected_data.apply(lambda x: x[i])
 
     return df
 
@@ -64,7 +67,7 @@ def create_plot(dfs:list[pd.DataFrame], ids:list[str], interval:list[int, int]=N
             df = df.iloc[interval[0]:interval[1]]
         for i, column in enumerate(df.columns[1:]):
             fig.add_trace(go.Scatter(
-                x=df[df.columns[0]], 
+                x=list(range(len(df[df.columns[0]]))), 
                 y=df[column], 
                 mode='lines', 
                 name=f"item_{df_id} - {column}",
@@ -120,7 +123,7 @@ def create_statistic(dfs: list[pd.DataFrame], ids: list[str], interval:list[int,
     combined_stats_df = combined_stats_df.applymap(lambda x: round(x, 2) if isinstance(x, (int, float)) else x)
     return combined_stats_df
 
-def clean_up_df(df: pd.DataFrame, rows_to_include: list[int]) -> pd.DataFrame:
+def clean_up_df(df: pd.DataFrame, rows_to_include: list[int], name_mapping_map:dict|None=None) -> pd.DataFrame:
     """
     清理数据集，将嵌套的np.ndarray列展平为多列。
     """
@@ -131,7 +134,7 @@ def clean_up_df(df: pd.DataFrame, rows_to_include: list[int]) -> pd.DataFrame:
         periods=len(row['target'][0]) if isinstance(row['target'][0], np.ndarray) else len(row['target']), 
         freq=row['freq']
     ).to_pydatetime().tolist(), axis=1)
-    df = flatten_ndarray_column(df, 'target', rows_to_include)
+    df = flatten_ndarray_column(df, 'target', rows_to_include, name_mapping_map)
     # 删除原始的start和freq列
     df.drop(columns=['start', 'freq', 'target'], inplace=True)
     if 'past_feat_dynamic_real' in df.columns: