[Enhancement]: Slightly more flexible API for add_missing_bounds

[pochedls]

Is your feature request related to a problem?

This is a convenience request. I almost always need to add missing bounds for one axis, but the (updated?) API wants a list.

Describe the solution you'd like

I'd like the API to accept a list or string (e.g., ["X", "Y"] or simply T).

Describe alternatives you've considered

No response

Additional context

No response

[lee1043]

I like this idea!

[tomvothecoder]

Sounds good to me. These APIs also need to be updated:

• xcdat/xcdat/dataset.py

  Lines 44 to 47 in c653bf2

  	def open_dataset(
  	path: str | os.PathLike[Any] | BufferedIOBase | AbstractDataStore,
  	data_var: Optional[str] = None,
  	add_bounds: List[CFAxisKey] | None = ["X", "Y"],

• xcdat/xcdat/dataset.py

  Lines 127 to 130 in c653bf2

  	def open_mfdataset(
  	paths: str | NestedSequence[str | os.PathLike],
  	data_var: Optional[str] = None,
  	add_bounds: List[CFAxisKey] | None = ["X", "Y"],

• xcdat/xcdat/dataset.py

  Lines 463 to 467 in c653bf2

  	def _postprocess_dataset(
  	dataset: xr.Dataset,
  	data_var: Optional[str] = None,
  	center_times: bool = False,
  	add_bounds: List[CFAxisKey] | None | bool = ["X", "Y"],

[pochedls]

Hi @tomvothecoder – I'm not opposed to updating the API in these places, but does it need to be updated there?

When opening a dataset it kind of makes sense to pass in a list for most use cases (make sure this has time/lat/lon bounds no matter what...). A lot of times when you add_missing_bounds it is because you got an error and now need to fix it by adding bounds for a specific axis.

[tomvothecoder]

When opening a dataset it kind of makes sense to pass in a list for most use cases (make sure this has time/lat/lon bounds no matter what...). A lot of times when you add_missing_bounds it is because you got an error and now need to fix it by adding bounds for a specific axis.

That's true. I was thinking of consistency across APIs but the use cases aren't exactly the same. We can keep those APIs as is.

[pochedls]

This is currently working on 0.7.1 (even though the function signature documents an argument of form axes : List[str]). I'm pretty sure this did not work in the past, but it is working for me now (I should have put in a minimal example...). I think this is because if I iterate over ['X', 'Y'] or X both return a string ("X" and "Y" versus "X", respectively). We can re-open this if stops working.

An example (that works):
wget https://www.ncei.noaa.gov/data/noaa-global-surface-temperature/v6/access/gridded/NOAAGlobalTemp_v6.0.0_gridded_s185001_e202405_c20240608T152402.nc (or a recent dataset here)

import xcdat as xc
fn = '/p/user_pub/climate_work/pochedley1/surface/NOAAGlobalTemp_v6.0.0_gridded_s185001_e202405_c20240608T152402.nc'
ds = xc.open_dataset(fn)
ds.bounds.get_bounds('T')

KeyError: "No bounds data variables were found for the 'T' axis. Make sure the dataset has bound data vars and their names match the 'bounds' attributes found on their related time coordinate variables. Alternatively, you can add bounds with ds.bounds.add_missing_bounds() or ds.bounds.add_bounds()."

ds = ds.bounds.add_missing_bounds('T')
ds.bounds.get_bounds('T')

<xarray.DataArray 'time_bnds' (time: 2093, bnds: 2)> Size: 33kB
array([[cftime.DatetimeGregorian(1850, 1, 1, 0, 0, 0, 0, has_year_zero=False),
cftime.DatetimeGregorian(1850, 2, 1, 0, 0, 0, 0, has_year_zero=False)],
...
[cftime.DatetimeGregorian(2024, 5, 1, 0, 0, 0, 0, has_year_zero=False),
cftime.DatetimeGregorian(2024, 6, 1, 0, 0, 0, 0, has_year_zero=False)]],
dtype=object)
Coordinates:

• time (time) object 17kB 1850-01-01 00:00:00 ... 2024-05-01 00:00:00
  Dimensions without coordinates: bnds
  Attributes:
  xcdat_bounds: True

[tomvothecoder]

This is currently working on 0.7.1 (even though the function signature documents an argument of form axes : List[str]). I'm pretty sure this did not work in the past, but it is working for me now (I should have put in a minimal example...). I think this is because if I iterate over ['X', 'Y'] or X both return a string ("X" and "Y" versus "X", respectively). We can re-open this if stops working.

The for loop below is supposed to loop over a list of strings. However, if we pass a single string, it'll loop over each character of the string. In your case you pass a single character string (e.g., "T") which means the for-loop behaves correctly (unintentionally).

xcdat/xcdat/bounds.py

Lines 166 to 170 in 7751452

	for axis in axes:
	try:
	coords = get_dim_coords(ds, axis)
	except KeyError:
	continue

We should do the following:

1. Update the docstring with axes: List[CFAxisKey] | CFAxisKEy
2. Add a check to make sure the single character is a CFAxisKey and raise a clear ValueError if it isn't