"""Count tornadoes by month for given states in the SPC 1950-2024 tornado CSV (argument 1).
Counts one record per tornado (sg == 1), assigned to the state where the record places it (st).
Run with python3 -I -X utf8."""
import sys, csv, collections

path = sys.argv[1]
states = sys.argv[2:] or ['AZ', 'NM']
by_month = {s: collections.Counter() for s in states}
years_oct = {s: collections.Counter() for s in states}
oct_days = {s: set() for s in states}
oct_mag = {s: collections.Counter() for s in states}
total = {s: 0 for s in states}
n_rows = 0
with open(path, newline='', encoding='utf-8', errors='replace') as f:
    r = csv.DictReader(f)
    cols = r.fieldnames
    for row in r:
        n_rows += 1
        if row['sg'].strip() != '1':
            continue
        st = row['st'].strip()
        if st not in by_month:
            continue
        mo = int(row['mo'])
        by_month[st][mo] += 1
        total[st] += 1
        if mo == 10:
            years_oct[st][int(row['yr'])] += 1
            oct_days[st].add(row['date'])
            oct_mag[st][row['mag'].strip()] += 1
print('columns:', cols)
print('rows:', n_rows)
for s in states:
    print('\n==', s, 'tornadoes 1950-2024 (sg=1):', total[s])
    print('  by month:', [(m, by_month[s][m]) for m in range(1, 13)])
    print('  October:', by_month[s][10], '= %.1f%% of all' % (100.0 * by_month[s][10] / total[s]))
    print('  years with an October tornado:', len(years_oct[s]), 'of 75;', sorted(years_oct[s].items()))
    print('  October tornado days:', len(oct_days[s]))
    print('  October ratings (mag):', sorted(oct_mag[s].items()))
    print('  average per year, all months: %.1f; October average: %.2f' % (total[s] / 75.0, by_month[s][10] / 75.0))
    last30 = sum(1 for y in years_oct[s] if y >= 1995)
    print('  1995-2024: October tornadoes', sum(c for y, c in years_oct[s].items() if y >= 1995), 'in', last30, 'years')