wsgi.py

@@ -8,7 +8,7 @@ import matplotlib.pyplot as plt
 import matplotlib.ticker as mtick
 import numpy as np
 import pandas as pd
-from flask import Flask, Markup, render_template, request
+from flask import Flask, render_template, request
 from flask_caching import Cache
 
 from download_digital import construct_dataframe, get_bez_data
@@ -48,10 +48,9 @@ def plot(
     data_folder: str = "data",
     sheet_name: str = "digital",
     total_target: int = 1500,
+    plot_all: bool = False,
     alpha: float | None = None,
 ) -> str:
-    curr_datetime = datetime.datetime.now()
-
     data_dict = {}
 
     ## Important: If multiple results are stored for the same date
@@ -74,22 +73,22 @@ def plot(
 
     df = series.to_frame("Digitale Befragung")
     df = df.reindex(
-        pd.date_range(start="2023-08-15", end=curr_datetime)
+        pd.date_range(start="2023-08-15", end=max(data_dict.keys()))
         + pd.DateOffset(hours=10)
     )
 
     if current_df is not None:
         current_df = current_df.astype({"Digitale Befragung": "Int32"})
         sum_val = current_df[["Digitale Befragung"]].sum().iloc[0]
-        df.loc[curr_datetime] = sum_val
+        df.loc[datetime.datetime.now()] = sum_val
-
-        if pd.isna(df.loc[df.index.max()][0]):
-            df = df.drop([df.index.max()])
 
     fig = plt.figure(dpi=300)
 
     # fill weekends
-    max_date = curr_datetime + datetime.timedelta(days=1)
+    max_date = max(data_dict.keys())
+    max_date = datetime.datetime.strptime(max_date, "%Y-%m-%d") + datetime.timedelta(
+        days=1
+    )
     days = pd.date_range(start="2023-08-14", end=max_date)
     for idx, day in enumerate(days[:-1]):
         if day.weekday() >= 5:
@@ -167,6 +166,16 @@ def plot(
 def tables(
     url: str = "https://beschaeftigtenbefragung.verdi.de/",
 ):
+    df, df_state = get_tables(url)
+
+    df = df.sort_values(
+        ["Digitale Befragung", "Bundesland", "Bezirk"], ascending=[False, True, True]
+    )
+
+    df_state = df_state.sort_values("Landesbezirk")
+
+    output_str = []
+
     def _print_as_html(df: pd.DataFrame):
         df = df.astype({"Digitale Befragung": "Int32"})
         with pd.option_context("display.max_rows", None):
@@ -197,44 +206,15 @@ def tables(
             output_str.append(tfoot)
             output_str.append(table[idx:])
 
-    output_str = []
-
-    try:
-        df, df_state = get_tables(url)
-
-        df = df.sort_values(
-            ["Digitale Befragung", "Bundesland", "Bezirk"],
-            ascending=[False, True, True],
-        )
-
-        df_state = df_state.sort_values("Landesbezirk")
-
-        image = plot(df_state)
-        timestamp = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-
-    except Exception:
-        last_file = sorted(Path("data").iterdir())[-1]
-        key = last_file.name[:10]
-
-        with (Path("data") / f"{key}_data.ods").open("rb") as ff:
-            df = pd.read_excel(ff, sheet_name="digital").astype(
-                {"Digitale Befragung": "Int32"}
-            )
-        with (Path("data") / f"{key}_state_data.ods").open("rb") as ff:
-            df_state = pd.read_excel(ff, sheet_name="digital").astype(
-                {"Digitale Befragung": "Int32"}
-            )
-
-        image = plot()
-        timestamp = Markup(f'<font color="red">{key} 10:00:00</font>')
-
     _print_as_html(df_state)
     _print_as_html(df)
 
+    image = plot(df_state, plot_all=True)
+
     return render_template(
         "base.html",
         tables="\n".join(output_str),
-        timestamp=timestamp,
+        timestamp=datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
         image=image,
     )