prevent calls to uds.grid in dfmt

[veenstrajelmer]

remove_ghostcells fails for grid with langeduikers. This raises "TypeError: Can only access grid topology via '.grid' if dataset contains exactly one grid. Dataset contains 2 grids. Use '.grids' instead."

When setting remove_edges=False and remove_ghost=False and using xugrid>=0.9.0, the dataset is successfully read. However, more errors will follow when using other dfm_tools functions due to .grid calls.

MWE:

import dfm_tools as dfmt
#file_nc = r'p:\11209233-mad09a-2023-rijn\C_Work\02_lange_duikers\02_berekeningen\dflowfm2d-rijn-j19_6-v2a\computations\hist\hw2018_kort\v15\results\*_fou.nc'
file_nc = r'p:\11209233-mad09a-2023-rijn\C_Work\02_lange_duikers\02_berekeningen\dflowfm2d-rijn-j19_6-v2a\computations\hist\hw2018_kort\v15\results\converted_RIJN_0*_map.nc'
# file_nc = r'p:\1230882-emodnet_hrsm\GTSMv5.0\SO_NHrivGTSM\UGgtsm_snipbd_v2_withriv_cellinfo_net.nc'
# file_nc = r"p:\1230882-emodnet_hrsm\GTSMv5.0\SO_NHrivGTSM\computations\BD014_fix_mapformat4\output\gtsm_model_0*_map.nc"
uds = dfmt.open_partitioned_dataset(file_nc)#, remove_edges=False, remove_ghost=False)

Prevent calls to .grid in dfmt code, or escape edges/ghost functions in case of multiple grids. This is at least the case in remove_ghostcells and remove_unassociated_edges. But there are more .grid calls in dfm_tools:

• xugrid_helpers.py: dimn_faces = grid.face_dimension, dimn_nodes = grid.node_dimension, dimn_edges = grid.edge_dimension, fill_value = grid.fill_value, indexer_np = grid.face_node_connectivity, gridname = uds.grid.name, uds.grid.to_dataset()[gridname], isinstance(uds.grid,xu.Ugrid1d), uds.grid.validate_edge_node_connectivity(), uds.grid.face_node_coordinates, uds.grid.meshkernel, uds.grid.crs, uds.grid.is_geographic
• get_nc.py: gridname = uds.grid.name, uds.grid.face_dimension, uds.grids[0].is_geographic, uds.grid.bounds.
• get_nc_helpers.py: gridname = ds.grid.name
• meshkernel_helpers.py: grid_is_geographic = uds.grid.is_geographic
• interpolate_grid2bnd.py: facedim = uds.grid.face_dimension