Parsing OFX datetime in Python

I'm trying to parse the datetime specified in the OFX 2.3 spec in Python. I believe it's a custom format, but feel free to let me know if it has a name. The spec states the following:

There is one format for representing dates, times, and time zones. The complete form is: YYYYMMDDHHMMSS.XXX [gmt offset[:tz name]]

For example, “19961005132200.124[-5:EST]” represents October 5, 1996, at 1:22 and 124 milliseconds p.m., in Eastern Standard Time. This is the same as 6:22 p.m. Greenwich Mean Time (GMT).

Here is my current attempt:

from datetime import datetime

date_str = "19961005132200.124[EST]"
date = datetime.strptime(date_str, "%Y%m%d%H%M%S.%f[%Z]")

This partial example works so far, but is lacking the GMT offset portion (the -5 in [-5:EST]). I'm not sure how to specify a time zone offset of at most two digits.

Some things to note here, first (as commented):

• Python built-in strptime will have a hard time here - %z won't parse a single digit offset hour, and %Z won't parse some (potentially) ambiguous time zone abbreviation.

Then, the OFX Banking Version 2.3 docs (sect. 3.2.8.2 Date and Datetime) leave some questions open to me:

• Is the UTC offset optional ?
• Why is EST called a time zone while it's just an abbreviation ?
• Why in the example the UTC offset is -5 hours while on 1996-10-05, US/Eastern was at UTC-4 ?
• What about offsets that have minutes specified, e.g. +5:30 for Asia/Calcutta ?
• (opinionated) Why re-invent the wheel in the first place instead of using a commonly used standard like ISO 8601 ?

Anyway, here's an attempt at a custom parser:

from datetime import datetime, timedelta, timezone
from zoneinfo import ZoneInfo

def parseOFXdatetime(s, tzinfos=None, _tz=None):
    """
    parse OFX datetime string to an aware Python datetime object.
    """
    # first, treat formats that have no UTC offset specified.
    if not '[' in s:
        # just make sure default format is satisfied by filling with zeros if needed
        s = s.ljust(14, '0') + '.000' if not '.' in s else s
        return datetime.strptime(s, "%Y%m%d%H%M%S.%f").replace(tzinfo=timezone.utc)

    # offset and tz are specified, so first get the date/time, offset and tzname components
    s, off = s.strip(']').split('[')
    off, name = off.split(':')
    s = s.ljust(14, '0') + '.000' if not '.' in s else s
    # if tzinfos are specified, map the tz name:
    if tzinfos:
        _tz = tzinfos.get(name) # this might still leave _tz as None...
    if not _tz: # ...so we derive a tz from a timedelta
        _tz = timezone(timedelta(hours=int(off)), name=name)
    return datetime.strptime(s, "%Y%m%d%H%M%S.%f").replace(tzinfo=_tz)


# some test strings

t = ["19961005132200.124[-5:EST]", "19961005132200.124", "199610051322", "19961005",
     "199610051322[-5:EST]", "19961005[-5:EST]"]

for s in t:
    print(# normal parsing
          f'{s}\n {repr(parseOFXdatetime(s))}\n'
          # parsing with tzinfo mapping supplied; abbreviation -> timezone object
          f' {repr(parseOFXdatetime(s, tzinfos={"EST": ZoneInfo("US/Eastern")}))}\n\n')