Aside from the plan to increase the speed (#122), while working on an MM CLI for @jordanschnell I found some things that should be fixed in the current OpenAQ reader that uses openaq-fetches.
[x] returned df currently may have some times outside of the requested dates (i.e. [greater than the ub](https://github.com (/noaa-oar-arl/monetio/blob/01e14d36d7b43e559a2f817d109657c0ddfc34a4/monetio/obs/openaq.py#L123)) (in #141)
[ ] returned df may have non-unique site IDs at given time.
Seems like there are two things going on here:
lat/lon that are only slightly different, the precision of the lat/lon in the current site IDs doesn't capture. Maybe create new site IDs, e.g. country code + hash of full-precision lat/lon (in #141)
lat/lon are the same but different measurements in different rows (e.g. 2 rows). Should combine these into one row where possible.
[x] typo in BC's new variable name with units (fixed in 01e14d3)
[x] add some tests which load a reduced set of files (in #141)
Aside from the plan to increase the speed (#122), while working on an MM CLI for @jordanschnell I found some things that should be fixed in the current OpenAQ reader that uses openaq-fetches.