Option to return error for cells without overlap

[nncarlson]

One argument to interpolate_mesh_var specifies how to handle empty cells, that is cells in the target mesh/part that don't intersect cells from the source mesh/part. In my application, such an occurrence indicates a fundamental problem with the pair of meshes or parts that I need to catch and handle (likely by aborting the simulation). Currently that doesn't appear possible.

What I'd like is for PartsPair::check_mesh to optionally return a status code for such an occurrence, or a new method to query the object about it.

An additional possibility (not well suited to my application) is to delay until interpolate_mesh_var with a new ERROR enum value for Empty_fuxup_type, that would trigger an error return.

[raovgarimella]

I think it's a good suggestion to have the possibility of Portage returning an indication that there were empty cells. Perhaps counts of partially_filled cells and empty_cells? Then an application can abort if they see that there are some empty cells. This should be easy to add. The other option that we have heard interest in is to fill with a background material. We will create a ticket for this. Until then the only options for you are EXTRAPOLATE and LEAVE_EMPTY.

…

-------------------------------------------------------------------- Rao V Garimella Tel: (505) 665-2928 T-5, MS B284 '\ '\ FAX: (505) 665-5757 Los Alamos National Lab ( )-( ) Email: [email protected] Los Alamos, NM 87545 Web: http://math.lanl.gov/~rao --------------------------------------------------------------------

________________________________ From: Neil Carlson <[email protected]> Sent: Wednesday, November 27, 2019 11:34:20 AM To: laristra/portage Cc: Subscribed Subject: [laristra/portage] Option to return error for cells without overlap (#29) One argument to interpolate_mesh_var specifies how to handle empty cells, that is cells in the target mesh/part that don't intersect cells from the source mesh/part. In my application, such an occurrence indicates a fundamental problem with the pair of meshes or parts that I need to catch and handle (likely by aborting the simulation). Currently that doesn't appear possible. What I'd like is for PartsPair::check_mesh to optionally return a status code for such an occurrence, or a new method to query the object about it. An additional possibility (not well suited to my application) is to delay until interpolate_mesh_var with a new ERROR enum value for Empty_fuxup_type, that would trigger an error return. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#29?email_source=notifications&email_token=AEBAZQD3KRODHU4RBJLRGODQV24SZA5CNFSM4JSKYW62YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4H4PXMDA>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AEBAZQB5YEWK7MTX7CTZGBLQV24SZANCNFSM4JSKYW6Q>.