plomlompom.com Git - berlin-corona-table/blob - enhance_table.py

   1 #!//usr/bin/env python3
   2
   3 # District population numbers as per Wikipedia.
   4 district_pops = {
   5   'CW': 342332,
   6   'FK': 289762,
   7   'Li': 291452,
   8   'MH': 268548,
   9   'Mi': 384172,
  10   'Ne': 329691,
  11   'Pa': 407765,
  12   'Re': 265225,
  13   'Sp': 243977,
  14   'SZ': 308697,
  15   'TS': 351644,
  16   'TK': 271153,
  17   'sum': 3754418,
  18 }
  19
  20 # Map abbreviations to full names.
  21 translate = {
  22   'CW': 'Charlottenburg-Wilmersdorf',
  23   'FK': 'Friedrichshain-Kreuzberg',
  24   'Li': 'Lichtenberg',
  25   'MH': 'Marzahn-Hellersdorf',
  26   'Mi': 'Mitte',
  27   'Ne': 'Neukölln',
  28   'Pa': 'Pankow',
  29   'Re': 'Reinickendorf',
  30   'Sp': 'Spandau',
  31   'SZ': 'Steglitz-Zehlendorf',
  32   'TS': 'Tempelhof-Schöneberg',
  33   'TK': 'Treptow-Köpenick',
  34   'sum': 'all of Berlin',
  35   'wsum': 'sum for last 7 days',
  36   'wavg': 'per-day average of new infections for last 7 days',
  37   'winc': 'incidence (x per 100k inhabitants) of new infections for last 7 days',
  38 }
  39
  40 # Read infections table path and output type.
  41 import sys
  42 if len(sys.argv) != 3:
  43     print('Expecting infections table file path and output type as only arguments.')
  44     exit(1)
  45 infections_table = sys.argv[1]
  46 output_type = sys.argv[2]
  47
  48 # Read infections table file lines.
  49 f = open(infections_table, 'r')
  50 lines = f.readlines()
  51 f.close()
  52
  53 # Basic input validation.
  54 import datetime
  55 header_elements = lines[0].split()
  56 if set(header_elements) != district_pops.keys() or \
  57        len(header_elements) != len(district_pops.keys()):
  58     raise Exception('infections table: invalid header')
  59 line_count = 0
  60 for line in lines[1:]:
  61     line_count += 1
  62     fields = line.split()
  63     if len(header_elements) != len(fields) - 1:
  64         raise Exception('infections table: too many elements on line %s',
  65                         line_count)
  66     try:
  67         datetime.date.fromisoformat(fields[0])
  68     except ValueError:
  69         raise Exception('infections table: bad ISO date on line %s',
  70                         line_count)
  71     for field in fields[1:]:
  72         try:
  73             int(field)
  74         except ValueError:
  75             raise Exception('infections table: bad value on line %s',
  76                             line_count)
  77
  78 # Parse first table file line for the names and order of districts.
  79 db = {}
  80 sorted_districts = []
  81 for header in lines[0].split():
  82     sorted_districts += [header]
  83     db[header] = {}
  84
  85 # Seed DB with daily new infections data per district, per date.
  86 sorted_dates = []
  87 for line in lines[1:]:
  88     fields = line.split()
  89     date = fields[0]
  90     sorted_dates += [date]
  91     for i in range(len(sorted_districts)):
  92         district = sorted_districts[i]
  93         district_data = fields[i + 1]
  94         db[district][date] = {'new_infections': int(district_data)}
  95 sorted_dates.sort()
  96
  97 # In LaGeSo's data, the last "district" is actually the sum of all districts /
  98 # the whole of Berlin.
  99 #
 100 # Fail on any day where the "sum" district's new infections are not the proper
 101 # sum of the individual districts new infections.  Yes, sometimes Lageso sends
 102 # data that is troubled in this way.  It will then have to be fixed manually in
 103 # the table file, since we should have a human look at what mistake was
 104 # probably made.
 105 for date in sorted_dates:
 106     sum_district = sorted_districts[-1]
 107     day_sum = 0
 108     for district in sorted_districts[:-1]:
 109         day_sum += db[district][date]['new_infections']
 110     if day_sum != db[sum_district][date]['new_infections']:
 111         raise Exception('Questionable district infection sum in %s' % date)
 112
 113 # Enhance DB with data about weekly sums, averages, incidences per day.  Ignore
 114 # days that have less than 6 predecessors (we can only know a weekly average if
 115 # we have a whole week of data).
 116 for i in range(len(sorted_dates)):
 117     if i < 6:
 118         continue
 119     date = sorted_dates[i]
 120     week_dates = []
 121     for j in range(7):
 122         week_dates += [sorted_dates[i - j]]
 123     for district in sorted_districts:
 124         district_pop = district_pops[district]
 125         week_sum = 0
 126         for week_date in week_dates:
 127             week_sum += db[district][week_date]['new_infections']
 128         db[district][date]['week_sum'] = week_sum
 129         db[district][date]['week_average'] = week_sum / 7
 130         db[district][date]['week_incidence'] = (week_sum / district_pop) * 100000
 131
 132 # Optimized for web browser viewing.
 133 if output_type == 'html':
 134     print("""<html>
 135 <style>
 136 table, tr, th, td { border: 1px solid black; }
 137 .day_row:nth-child(7n+2) { background-color: yellow; }
 138 .district_name { writing-mode: vertical-rl; transform: rotate(180deg); }
 139 </style>
 140 <h1>Table of Berlin's Corona infection number development by districts.
 141 </h1>
 142 <p>Updated daily at 9pm. <a href="https://plomlompom.com/repos/?p=berlin-corona-table">Source code</a>. <a href="berlin_corona.txt">Text view optimized for terminal curl</a>.</p>
 143 <table>
 144 <tr>
 145 <th>date</th>""")
 146     sorted_dates.reverse()
 147     sum_district = sorted_districts[-1]
 148     for district in sorted_districts:
 149         long_form = translate[district]
 150         if sum_district == district:
 151             print('<th>%s</th>' % long_form)
 152         else:
 153             print('<th class="district_name">%s</th>' % long_form)
 154     print('</tr>')
 155     for date in sorted_dates:
 156         print('<tr class="day_row">')
 157         print('<td>%s</td>' % date)
 158         long_wsum = translate['wsum']
 159         long_wavg = translate['wavg']
 160         long_winc = translate['winc']
 161         for district in sorted_districts:
 162             district_data = db[district][date]
 163             week_sum = week_avg = week_inc = ''
 164             new_infections = district_data['new_infections']
 165             if 'week_sum' in district_data:
 166                 week_sum = '%s' % district_data['week_sum']
 167             if 'week_average' in district_data:
 168                 week_avg = '%.1f' % district_data['week_average']
 169             if 'week_incidence' in district_data:
 170                 week_inc = '%.1f' % district_data['week_incidence']
 171             print('<td>')
 172             print(new_infections)
 173             if district == sum_district:
 174                 print('<details open><summary></summary><table>')
 175             else:
 176                 print('<details><summary></summary><table>')
 177             print('<tr><th>%s</th><td>%s</td></tr>' % (long_wsum, week_sum))
 178             print('<tr><th>%s</th><td>%s</td></tr>' % (long_wavg, week_avg))
 179             print('<tr><th>%s</th><td>%s</td></tr>' % (long_winc, week_inc))
 180             print('</table></details>')
 181             print('</td>')
 182         print('</tr>')
 183     print('</table>')
 184     print('</html>')
 185
 186 # Optimized for in-terminal curl.
 187 elif output_type == 'txt':
 188
 189     # Explain what this is.
 190     intro = \
 191 """Table of Berlin's Corona infection number development by districts.
 192 Updated daily around 9pm.
 193
 194 Abbrevations/explanations:
 195 """
 196     for k in translate:
 197         intro += "%s: %s\n" % (k, translate[k])
 198     intro += """
 199 Source code: https://plomlompom.com/repos/?p=berlin-corona-table
 200
 201 HTML view: https://plomlompom.com/berlin_corona.html
 202 """
 203     print(intro)
 204
 205     # Output table of enhanced daily infection data, newest on top,
 206     # separated into 7-day units.
 207     sorted_dates.reverse()
 208     weekday_count = 0
 209     sum_district = sorted_districts[-1]
 210     for date in sorted_dates:
 211
 212         # Week table header.
 213         if weekday_count == 0:
 214             print(' '*11, '  '.join(sorted_districts[:-1]),
 215                   sorted_districts[-1], 'wsum', ' wavg', 'winc')
 216             week_start_date = date
 217
 218         # Day data line.
 219         new_infections = []
 220         for district in sorted_districts:
 221             new_infections += [db[district][date]['new_infections']]
 222         week_sum = week_avg = week_inc = ''
 223         sum_district_data = db[sum_district][date]
 224         if 'week_sum' in sum_district_data:
 225             week_sum = '%4s' % sum_district_data['week_sum']
 226         if 'week_average' in sum_district_data:
 227             week_avg = '%5.1f' % sum_district_data['week_average']
 228         if 'week_incidence' in sum_district_data:
 229             week_inc = '%4.1f' % sum_district_data['week_incidence']
 230         print(date, ' '.join(['%3s' % infections
 231                               for infections in new_infections]),
 232               week_sum, week_avg, week_inc)
 233
 234         # Maintain 7-day cycle.
 235         weekday_count += 1
 236         if weekday_count != 7:
 237             continue
 238         weekday_count = 0
 239
 240         # After each 7 days, print summary for individual districts.
 241         weekly_sums = []
 242         weekly_avgs = []
 243         weekly_incs = []
 244         for district in sorted_districts[:-1]:
 245             weekly_sums += [db[district][week_start_date]['week_sum']]
 246             weekly_avgs += [db[district][week_start_date]['week_average']]
 247             weekly_incs += [db[district][week_start_date]['week_incidence']]
 248         print()
 249         print('district stats for week from %s to %s:' % (date, week_start_date))
 250         print(' '*7, '    '.join(sorted_districts[:-1]))
 251         print('wsum', ' '.join(['%5.1f' % wsum for wsum in weekly_sums]))
 252         print('wavg', ' '.join(['%5.1f' % wavg for wavg in weekly_avgs]))
 253         print('winc', ' '.join(['%5.1f' % winc for winc in weekly_incs]))
 254         print()