Months may be skipped in monthly metrics histories when last day of first month in series is greater than in a subsequent month

[bryanlandia]

Appears that we never get any kind of metrics for February (unless probably it's the current month). This affects all types of metrics.

Looks like the problem is in metrics.previous_months_iterator

>>> from datetime import datetime
>>> date_for = datetime.utcnow().date()
>>> from figures import helpers, metrics
>>> date_for = helpers.as_date(date_for)
>>> date_for
datetime.date(2021, 5, 31)
>>> metrics.previous_months_iterator(month_for=date_for, months_back=6)
<generator object previous_months_iterator at 0x7fbc30ec0190>
>>> [mon for mon in metrics.previous_months_iterator(month_for=date_for, months_back=6)]
[(2020, 11, 30), (2020, 12, 31), (2021, 1, 31), (2021, 3, 31), (2021, 4, 30), (2021, 5, 31)]

[bryanlandia]

I haven't yet verified if Feb's data is appearing as January or if February is just skipped altogether.

[bryanlandia]

See note here https://dateutil.readthedocs.io/en/stable/rrule.html#classes

Per RFC section 3.3.10, recurrence instances falling on invalid dates and times are ignored rather than coerced:

Recurrence rules may generate recurrence instances with an invalid date (e.g., February 30) or nonexistent local time (e.g., 1:30 AM on a day where the local time is moved forward by an hour at 1:00 AM). Such recurrence instances MUST be ignored and MUST NOT be counted as part of the recurrence set.

This can lead to possibly surprising behavior when, for example, the start date occurs at the end of the month:

from dateutil.rrule import rrule, MONTHLY
from datetime import datetime
start_date = datetime(2014, 12, 31)
list(rrule(freq=MONTHLY, count=4, dtstart=start_date))
... # doctest: +NORMALIZE_WHITESPACE
[datetime.datetime(2014, 12, 31, 0, 0),
datetime.datetime(2015, 1, 31, 0, 0),
datetime.datetime(2015, 3, 31, 0, 0),
datetime.datetime(2015, 5, 31, 0, 0)]

[bryanlandia]

See draft PR #358 for fix

[johnbaldwin]

Thanks for this, @bryanlandia !

[bryanlandia]

@johnbaldwin I think you probably saw this, but most likely the change in #360 could be removed once #358 is merged. Anyway, the problem onlly occurs in days 29, 30, and 31 of each month 🤣 so we have a little time to sort it out. (I need to write or update tests)