fail to "vectorize" velocity_from_position

[selipot]

I am attempting to apply velocity_from_position to xarray.DataArrays of lon, lat, and time. I have been following a tutorial for a similar situation. With the following ds Dataset:

ds.info()
xarray.Dataset {
dimensions:
	trajectory = 593297 ;
	obs = 1440 ;

variables:
	float64 time(trajectory, obs) ;
	float32 lat(trajectory, obs) ;
	float32 lon(trajectory, obs) ;
	int32 obs(obs) ;
	int64 trajectory(trajectory) ;
}

I can easily do:

u,v = velocity_from_position(ds.lon.isel(trajectory=0),ds.lat.isel(trajectory=0),ds.time.isel(trajectory=0))

or

u2,v2 = xr.apply_ufunc(
    velocity_from_position,
    ds.lon.isel(trajectory=0),
    ds.lat.isel(trajectory=0),
    ds.time.isel(trajectory=0),
    input_core_dims=[["obs"], ["obs"], ["obs"]],
    output_core_dims=[["obs"], ["obs"]],
    dask="allowed",
)

but the following fails:

u2,v2 = xr.apply_ufunc(
    velocity_from_position,  # first the function
    ds.lon.isel(trajectory=slice(0,10)),
    ds.lat.isel(trajectory=slice(0,10)),
    ds.time.isel(trajectory=slice(0,10)),
    input_core_dims=[["obs"], ["obs"], ["obs"]],
    output_core_dims=[["obs"],["obs"]],
    dask="allowed",
    vectorize=True,
)

and the bottom line of the error is

File ~/miniconda3/envs/research/lib/python3.10/site-packages/clouddrift/analysis.py:65, in velocity_from_position(x, y, time, coord_system, difference_scheme)
     57 # Compute dx, dy, and dt
     58 if difference_scheme == "forward":
     59 
     60     # All values except the ending boundary value are computed using the
   (...)
     63 
     64     # Time
---> 65     dt[:-1] = np.diff(time)
     66     dt[-1] = dt[-2]
     68     # Space

ValueError: could not broadcast input array from shape (10,1439) into shape (9,1440)

So yes, I get the error but I don't understand if the fix is to apply ufunc differently or make velocity_from_position more flexible?