An algorithm to search longest date intervals among a list of dates in Python (currently searches longest date intervals in months)

algo.py

from datetime import datetime
from typing import List


def solve(date_items: List[datetime]):
    """
    Get the longest date interval from the date_items list.

    :param date_items:
    :return:
    """

    # 1) Sort the items
    sorted_date_items = sorted(date_items)

    # 2) Convert our items to a dict as "[item_year]_[item_month]__[nth_of_year] => item"
    nth_dates = {}
    for item in sorted_date_items:
        new_year_day = datetime(year=item.year, month=1, day=1)
        nth_of_year = (item - new_year_day).days + 1

        key = f'{item.year}_{item.month}_{nth_of_year}'
        nth_dates[key] = item

    # 3) Iterate over items
    count_map = {}
    item_start = None
    counter = 1
    for key, item in nth_dates.items():
        current_year, current_month, current_nth_day = key.split('_')
        current_nth_day = int(current_nth_day)
        next_key = f'{current_year}_{current_month}_{current_nth_day + 1}'

        if next_key in nth_dates:
            if item_start is None:
                item_start = item

            counter += 1
        else:
            if item_start is None:
                item_start = item

            if counter >= 2:
                # add to our map
                map_key = f'{current_year}_{current_month}_{counter}'
                count_map[map_key] = {'item_start': item_start, 'item_end': item}
                # reset tmp variables
                item_start = None
                counter = 1

    # 4) Transform elements to a list with its counters
    by_counter = []
    for key, value in count_map.items():
        _, _, counter = key.split('_')
        counter = int(counter)
        by_counter.append({
            'counter': counter,
            'value': value,
        })

    # 5) Sort the list by element's counters
    by_counter_sorted = sorted(by_counter, key=lambda it: it['counter'])
    best_match = by_counter_sorted[-1]

    result = {
        'date_interval': best_match['value'],
        'consecutive_days': best_match['counter']
    }

    return {'count_map': count_map, 'result': result}


if __name__ == '__main__':
    our_dates = [
        datetime(2019, 2, 19),

        datetime(2019, 2, 21),
        datetime(2019, 2, 22),

        datetime(2019, 4, 22),
        datetime(2019, 4, 23),
        datetime(2019, 4, 24),

        datetime(2020, 4, 23),
        datetime(2020, 4, 24),

        datetime(2020, 6, 26),
        datetime(2020, 6, 27),
        datetime(2020, 6, 28),
        datetime(2020, 6, 29),
        datetime(2020, 6, 30),

        datetime(2020, 6, 1),
        datetime(2020, 6, 2),
        datetime(2020, 6, 3),
        datetime(2020, 6, 4),
        datetime(2020, 6, 5),
        datetime(2020, 6, 6),
        datetime(2020, 6, 7),

        datetime(2020, 7, 11),
        datetime(2020, 7, 12),
        datetime(2020, 7, 13),
    ]

    longest_interval = solve(our_dates)

    result_count_map = longest_interval['count_map']
    for k, v in result_count_map.items():
        item_s = v['item_start'].strftime('%Y-%m-%d')
        item_e = v['item_end'].strftime('%Y-%m-%d')

        print(k, v)

        print(item_s, item_e)

    print()
    print('---')
    print()

    res = longest_interval['result']
    print(res)

console.log

> python algo.py

2019_2_2 {'item_start': datetime.datetime(2019, 2, 19, 0, 0), 'item_end': datetime.datetime(2019, 2, 22, 0, 0)}
2019-02-19 2019-02-22
2019_4_3 {'item_start': datetime.datetime(2019, 4, 22, 0, 0), 'item_end': datetime.datetime(2019, 4, 24, 0, 0)}
2019-04-22 2019-04-24
2020_4_2 {'item_start': datetime.datetime(2020, 4, 23, 0, 0), 'item_end': datetime.datetime(2020, 4, 24, 0, 0)}
2020-04-23 2020-04-24
2020_6_7 {'item_start': datetime.datetime(2020, 6, 1, 0, 0), 'item_end': datetime.datetime(2020, 6, 7, 0, 0)}
2020-06-01 2020-06-07
2020_6_5 {'item_start': datetime.datetime(2020, 6, 26, 0, 0), 'item_end': datetime.datetime(2020, 6, 30, 0, 0)}
2020-06-26 2020-06-30
2020_7_3 {'item_start': datetime.datetime(2020, 7, 11, 0, 0), 'item_end': datetime.datetime(2020, 7, 13, 0, 0)}
2020-07-11 2020-07-13

---

{'date_interval': {'item_start': datetime.datetime(2020, 6, 1, 0, 0), 'item_end': datetime.datetime(2020, 6, 7, 0, 0)}, 'consecutive_days': 7}