assume UTC, and drop timezone

[scottyhq]

simple fix for #2

in brief, currently xarray does not like timezone information that can be present in (pandas.DatetimeIndex(['2019-12-30 00:00:01+00:00'], dtype='datetime64[ns, UTC]', freq=None)). As a result nicely formated datetimes are converted to nanosecond integers, causing confusion for users.

STAC datetimes should all be UTC. This additional line ensures the UTC timezone info is present (due to a pandas bug with infer_datetime_format=True), and then drops the timezone info for xarray.

A few examples of what this does:

# want to ensure these work
print(pd.to_datetime('2018-10-27T23:30:00Z').tz_convert('UTC').tz_localize(None))
print(pd.to_datetime('2018-10-27T23:30:00.123Z').tz_convert('UTC').tz_localize(None))

# valid ISO8601, but unclear if found in wild STACs
print(pd.to_datetime('2018-10-28T01:30:00+02:00').tz_convert('UTC').tz_localize(None))

# an invalid ISO8601 string (no Z) -> probably ok to hit errors in this case
pd.to_datetime('2018-10-28 01:30:00.1').tz_convert('UTC').tz_localize(None)

2018-10-27 23:30:00
2018-10-27 23:30:00.123000
2018-10-27 23:30:00
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-21-3e456a4822cb> in <module>
      7 
      8 # an invalid ISO8601 string (no Z) -> probably ok to hit errors in this case
----> 9 pd.to_datetime('2018-10-28 01:30:00.1').tz_convert('UTC').tz_localize(None)

pandas/_libs/tslibs/timestamps.pyx in pandas._libs.tslibs.timestamps.Timestamp.tz_convert()

TypeError: Cannot convert tz-naive Timestamp, use tz_localize to localize

[gjoseph92]

Awesome, thanks for the fix @scottyhq!

Would you mind also adding a note about this in the stackstac.stack docstring? Maybe something in the returns section about the time coordinate always being converted to UTC?

[gjoseph92]

Excellent! Thank you!