Incorrect timestamp index description in iotools.era5 docstring

[williamhobbs]

Describe the bug
The ERA5 docstring says that the index (timestamp) correspond to the beginning of the interval, but as far as I can tell, it is the end of the interval.

pvlib-python/pvlib/iotools/era5.py

Line 105 in a53e022

Time series data. The index corresponds to the start of the interval.

Additional context
It isn't easy to find a clear description of ERA5 timestamps. See PyPSA/atlite#158 (comment), https://confluence.ecmwf.int/plugins/viewsource/viewpagesrc.action?pageId=85402030.

[AdamRJensen]

I recently compared measurement data from a BSRN station with ERA5 and can confirm that the ERA5 data returned by pvlib.iotools.get_era5 is labeled by the end of the interval.

[AdamRJensen]

I think this documentation page corroborates it, yet it is still quite poorly documented.

[williamhobbs]

I think this documentation page corroborates it, yet it is still quite poorly documented.

Agreed. That’s the best official source I could find. But still not great.