added nested pie chart

app.py

@@ -103,8 +103,10 @@ def main(from_date,
             for col in PLOT_CHOICES_REVERSE_DICT:  # plot_choices:
                 if col == 'merged_target':
                     pie_charts.append(bar(db[0], col + ": " + db[1]))
-                else:
+                elif col == 'sentiment':
                     pie_charts.append(pie_chart(db[0], col, col + ": " + db[1]))
+                elif col == TOPIC:
+                    pie_charts.append(nested_pie_chart(db[0], col, col + ": " + db[1]))
         return pie_charts
 
     def bar(db: pd.DataFrame, title):
@@ -162,7 +164,7 @@ def main(from_date,
         if db.empty:
             return None
         else:
-            # db = db[col_name].value_counts()[:5] # Lägg till "Others sedan"
+            # df = df[col_name].value_counts()[:5] # Lägg till "Others sedan"
             db = pie_chart_input(db, col_name, LIMIT)
             labels = db[col_name].to_list()
             sizes = db['frequency'].values
@@ -173,6 +175,48 @@ def main(from_date,
             plt.title(title, fontdict=font1)
             return fig
 
+    def nested_pie_chart(df, col_name, title):
+        if df.empty:
+            return None
+        else:
+            count_dict = {}
+            sent_dict = {'positive': 0, 'negative': 1, 'neutral': 2, 'other': 3}
+            tot_sum = len(df)
+            for i in range(df.shape[0]):
+                topic = df.iloc[i][TOPIC]
+                sentiment = df.iloc[i]['sentiment'] if df.iloc[i]['sentiment'] in sent_dict else 'other'
+                if topic not in count_dict:
+                    count_dict[topic] = [0, 0, 0, 0]
+                    count_dict[topic][sent_dict[sentiment]] += 1
+                else:
+                    count_dict[topic][sent_dict[sentiment]] += 1
+            count_list = []
+            other_list = np.array([0, 0, 0, 0])
+            labels = []
+            for topic in count_dict:
+                if tot_sum > 0 and np.sum(count_dict[topic]) / tot_sum > LIMIT:
+                    count_list.append(count_dict[topic])
+                    labels.append(topic)
+                else:
+                    other_list += np.array(count_dict[topic])
+
+            count_list.append(list(other_list))
+            labels.append('Other')
+            fig, ax = plt.subplots()
+
+            size = 0.3
+            vals = np.array(count_list)
+            inner_colors = ['green', 'red', 'blue', 'yellow'] * len(count_dict)
+            if vals.shape[0] == 0:
+                pass
+            else:
+                ax.pie(vals.sum(axis=1), radius=1, labels=labels, pctdistance=0.9,
+                       wedgeprops=dict(width=size, edgecolor='w'))
+                ax.pie(vals.flatten(), radius=1 - size, colors=inner_colors,
+                       wedgeprops=dict(width=size, edgecolor='w'))
+                ax.set(aspect='equal', title='Nested sentiment plot')
+            return fig
+
     # text_classifier = tc.TextClassifier(from_date=from_date, to_date=to_date,
     #                                     user_list=match_name_lower_case(usr_name_choices),
     #                                     num_tweets=NUM_TWEETS)
@@ -188,9 +232,9 @@ def main(from_date,
     # Remove entries from df where 'tweet' starts with '@'
     df = df[df['tweet'].str.startswith('@') == False]
     # change 'merged_topic' to 'Other' if it is 'ERROR_9000' or 'ERROR_496'
-    df['merged_topic'] = df['merged_topic'].apply(lambda x: "other" if x == "ERROR_9000" or x == "ERROR_496" else x)
+    df[TOPIC] = df[TOPIC].apply(lambda x: "N/A" if x == "ERROR_9000" or x == "ERROR_496" else x)
     # change 'merged_topic' to 'Government' if it is 's'
-    df['merged_topic'] = df['merged_topic'].apply(lambda x: "The Government" if x == "s" else x)
+    df[TOPIC] = df[TOPIC].apply(lambda x: "The Government" if x == "s" else x)
     if save_selected:
         user_list = match_name_lower_case(usr_name_choices)
         df_l = []
@@ -204,7 +248,7 @@ def main(from_date,
     else:
         save_selected_checkbox = [gr.Checkbox.update(interactive=True)]
 
-    pycharts = add_pie_chart(df, usr_name_choices, convert_plot_choices(plot_choice))
+    pie_charts = add_pie_chart(df, usr_name_choices, convert_plot_choices(plot_choice))
 
     rb_components = [rb1, rb2, rb3, rb4, rb5, rb6, rb7, rb8]  # radio_buttons
     df_visibility_check = [v1, v2, v3, v4, v5, v6, v7, v8]
@@ -232,8 +276,8 @@ def main(from_date,
 
         return df_list + number_tweets + save_file_components_list
 
-    return pycharts + save_selected_checkbox + get_selected_df_list(df, save_file_bool, list(usr_name_choices),
-                                                                    rb_components, df_visibility_check)
+    return pie_charts + save_selected_checkbox + get_selected_df_list(df, save_file_bool, list(usr_name_choices),
+                                                                      rb_components, df_visibility_check)
 
 
 ''' END OF MAIN 
@@ -285,8 +329,6 @@ def export_to_download(_data_frame, _type: str):
 
 
 def pie_chart_input(df, column, limit):
-    df_len = len(df)
-    df_v = df[column].value_counts()
     df_len = len(df)
     if column == "sentiment":
         ds_sentiment = df[column].apply(lambda x: re.sub("\s+", "", str(x)))
@@ -319,7 +361,7 @@ def pie_chart_input(df, column, limit):
     else:
         ind_other = freq_dict[column].index("other")
         freq_dict["frequency"][ind_other] += freq_other
-
+    test_frame = pd.DataFrame.from_dict(freq_dict)
     return pd.DataFrame.from_dict(freq_dict)