open_dataset(area=...) not working as expected

[icedoom888]

Describe the bug
When opening a dataset providing the specific (north, west, south, east) coordinates, the resulting dataset has different bounds: longitude alway ranges between 0 and 359.

** Version number **
-e git+ssh://[email protected]/MeteoSwiss-APN/anemoi-datasets.git@7cd397e6d4e9f58f923fcfc75cabe6eca289d1bf#egg=anemoi_datasets
-e git+ssh://[email protected]/MeteoSwiss/anemoi-core.git@bdc05d85458b50640979ffae2503b379d2202c9b#egg=anemoi_graphs&subdirectory=graphs
anemoi-inference==0.4.4
-e git+ssh://[email protected]/MeteoSwiss/anemoi-core.git@bdc05d85458b50640979ffae2503b379d2202c9b#egg=anemoi_models&subdirectory=models
-e git+ssh://[email protected]/MeteoSwiss/anemoi-core.git@bdc05d85458b50640979ffae2503b379d2202c9b#egg=anemoi_training&subdirectory=training
anemoi-transform==0.1.0
anemoi-utils==0.4.10

To Reproduce
global_name = "/scratch/mch/apennino/data/aifs-ea-an-oper-0001-mars-n320-1979-2022-6h-v6.zarr"
lam_name = "/scratch/mch/fzanetta/data/anemoi/datasets/mch-co2-an-archive-0p02-2015-2020-6h-v2-pl13.zarr"

selected_vars = [ "10u", "10v", "2d", "2t", "cos_julian_day", "cos_latitude", "cos_local_time", "cos_longitude", "insolation", "lsm", "msl", "q_50", "q_100", "q_150", "q_200", "q_250", "q_300", "q_400", "q_500", "q_600", "q_700", "q_850", "q_925", "q_1000", "sin_julian_day", "sin_latitude", "sin_local_time", "sin_longitude", "sp", "tp", "t_50", "t_100", "t_150", "t_200", "t_250", "t_300", "t_400", "t_500", "t_600", "t_700", "t_850", "t_925", "t_1000", "u_50", "u_100", "u_150", "u_200", "u_250", "u_300", "u_400", "u_500", "u_600", "u_700", "u_850", "u_925", "u_1000", "v_50", "v_100", "v_150", "v_200", "v_250", "v_300", "v_400", "v_500", "v_600", "v_700", "v_850", "v_925", "v_1000", "w_100", "w_1000", "w_150", "w_200", "w_250", "w_300", "w_400", "w_50", "w_500", "w_600", "w_700", "w_850", "w_925", "z", "z_100", "z_1000", "z_150", "z_200", "z_250", "z_300", "z_400", "z_50", "z_500", "z_600", "z_700", "z_850", "z_925"]

lam_dataset = open_dataset(lam_name, select=selected_vars)
l_coords = max(lam_dataset.latitudes), min(lam_dataset.longitudes), min(lam_dataset.latitudes), max(lam_dataset.longitudes)
print("LAM (north, west, south, east): ", l_coords)

Open global and cut it

lam_start = lam_dataset.dates[0]
lam_end = lam_dataset.dates[-1]

global_dataset = open_dataset(global_name, start=lam_start, end=lam_end, area=l_coords, select=selected_vars)
g_coords = max(global_dataset.latitudes), min(global_dataset.longitudes), min(global_dataset.latitudes), max(global_dataset.longitudes)
print("global-lam (north, west, south, east): ", g_coords)

URL to sample input data
Use any ERA5 and LAM dataset

Expected behavior
l_coords should be similar to g_coords

Screenshots
The LAM areas are also slightly off: