some more minor changes

app.py

@@ -1,17 +1,16 @@
 import numpy as np
-
+from matplotlib import pyplot as plt, use as plt_use
 from textclassifier import TextClassifier as tc
 import pandas as pd
 import regex as re
-
 from pathlib import Path
 import glob
-import matplotlib
+import os
 
-matplotlib.use('Agg')
-import matplotlib.pyplot as plt
-from functions import functions as f
-import time
+plt_use('Agg')
+
+# from functions import functions as f
+# import time
 
 SELECTED_COLUMN_DICT = {
     'merged_topic': ['tweet', 'main_topic', 'sub_topic', 'synonym_topic', 'cos_sim_topic', 'merged_topic'],
@@ -37,7 +36,7 @@ UserNameDict = dict(zip(['Jimmie Åkesson', 'Ebba Busch', 'Annie Lööf', 'Johan
 
 Columns = ['username', 'nlikes', 'nreplies', 'nretweets', 'main_topic', 'sub_topic', 'sentiment', 'target', 'tweet',
            'date', 'urls', 'id', 'class_tuple', 'user_id']
-num_tweet = 1000
+NUM_TWEETS = 1000
 LIMIT = 0.05
 
 
@@ -172,14 +171,33 @@ def main(from_date,
             plt.title(title, fontdict=font1)
             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_tweet)
-    text_classifier.run_main_pipeline()
-    dataframe = text_classifier.get_dataframe()
-    # dataframe= pd.read_csv(os.path.dirname(
-    #      os.path.dirname(os.path.abspath(__file__))) + "/politweet/data/twitterdata.csv")  #
-    df = dataframe
+    # Define a function that gives the proportion of sentiments as a function of date
+    def sentiment_date(df, col_name, title):
+        if df.empty:
+            return None
+        else:
+            df = sentiment_date_input(df, col_name, LIMIT)
+            labels = df[col_name].to_list()
+            sizes = df['frequency'].values
+            # explode = (0, 0.1, 0, 0, 0)  # only "explode" the 2nd slice (i.e. 'Hogs')
+            font1 = {'family': 'serif', 'color': 'blue', 'size': 20}
+            fig = plt.figure()
+            plt.pie(sizes, labels=labels, radius=1, autopct='%1.1f%%')
+            plt.title(title, fontdict=font1)
+            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)
+    # text_classifier.run_main_pipeline()
+    # dataframe = text_classifier.get_dataframe()
 
+    dataframe = pd.read_csv("{}/data/twitterdata.csv".format(tc.ROOT_PATH))
+    # choose subset between from_date and to_date and username is in usr_name_choices
+    df = dataframe.loc[(dataframe['date'] >= from_date) & (dataframe['date'] <= to_date) & \
+                              (dataframe['username'].isin(match_name_lower_case(usr_name_choices)))].copy()
+    # Sort df by date
+    df.sort_values(by=['date'], inplace=True)
     if save_selected:
         user_list = match_name_lower_case(usr_name_choices)
         df_l = []