cmip6_fix_fgoals_f3_l_Amon_time_bnds #831
Conversation
|
I would tackle the codacy issues. Though there is not much I can do, any ideas. |
| for cube in cubes: | ||
| if cube.attributes['table_id'] == 'Amon': | ||
| time = cube.coord('time') | ||
| if cube.attributes['table_id'] == 'Amon' and \ |
There was a problem hiding this comment.
you are checking for this already two lines up
| cftime.DatetimeNoLeap(c.year + 1, 1, 1) for c in times | ||
| ] | ||
| time.bounds = time.units.date2num( | ||
| np.stack([starts, ends], -1)) |
There was a problem hiding this comment.
is the fault discoverable immediately by looking at the first and second [b1, b2] bounds? Or could be somewhere deeper in the bounds array? I also think it's prob best if you just say time.bounds = None; time.guess_bounds() - bit more general?
There was a problem hiding this comment.
From what I have seen it's concerning all time bounds. I added a check over the whole period though.
Concerning boundary guessing, I don't consider it the way forward, as it simple finds bounds in the middle between the two points, which is not correct in the case of monthly data. Leading to e.g. DimCoord([1850-01-16 12:00:00, 1850-02-15 00:00:00, 1850-03-16 12:00:00 ...] , bounds=[[1850-01-01 18:00:00, 1850-01-31 06:00:00], [1850-01-31 06:00:00, 1850-03-01 18:00:00], [1850-03-01 18:00:00, 1850-03-31 18:00:00], ...]
There was a problem hiding this comment.
OK - am not a big fan of guessing bounds myself but it may be easier if the result is the same as the coded by hand one, which is not the case here 👍
Co-authored-by: Valeriu Predoi <valeriu.predoi@gmail.com>
|
@valeriupredoi same here I assume ;) |
Tasks
Closes #821