Time zone capital or lowercase

[scotthavens]

The time_zone value in the config file has to specifically be from the pytz list with correct captial case. For example, mst does not work but MST does. However, utc works when it should be UTC so somewhere SMRF is modifying the value.

https://stackoverflow.com/questions/13866926/is-there-a-list-of-pytz-timezones

[scotthavens]

While PR #145 address a lot of the timezone handling, inicheck is parsing the time_zone to lower. The only way around at the moment is to have inicheck load the config file, change the time_zone then pass the UserConfig object to run_smrf.

The solution to this would be to have inicheck type to not change the string at all.

[scotthavens]

There is a RawString field in katana that solved this problem.

https://github.com/USDA-ARS-NWRC/katana/blob/6a408cbf0b88aadf98a4ef8c58bd50901936f0ce/katana/CoreConfig.ini#L137

[jomey]

Alternatively you could use a little filter to match the input string against the available from ptyz

Example for 'mst' as given string, will return 'MST'

import re

pattern = re.compile('^mst$', re.I)
next(filter(pattern.match, pytz.all_timezones))

[scotthavens]

Would this make users lazy in not looking up the actual timezone name? Just passing the string through is working now with minimal changes. The gotcha is that pytz has both a utc and UTC which is why it's worked so far. However, if the tests are changed to UTC all the gold files will have to be changed as well just for the time zone attribute...

[jomey]

I am mainly suggesting a case-insensitive lookup of the user input. Has the additional advantage that it is checked against valid ones and can catch errors early

[scotthavens]

Fully understand. We need to figure out what is better, case insensitive or not?

[jomey]

You make a good point regards enforcing/educating users to lookup proper timezone names.
The doc can always refer to the accepted list: https://en.wikipedia.org/wiki/List_of_tz_database_time_zones

[jomey]

For down the road: USDA-ARS-NWRC/inicheck/issues/53