Issue 1810 - Attenuation property IO#1986
Conversation
|
Retest this please |
| @@ -122,7 +130,8 @@ int specfem::io::impl::write_medium_group( | |||
| const int n_elements = element_indices.size(); | |||
|
|
|||
| using DomainView3d = | |||
There was a problem hiding this comment.
This is wrong. It should be derived from data containers.
There was a problem hiding this comment.
This is a miss on my end. Cannot tell you how many times I interrupted claude to please not edit the medium_writer. I should have checked at the end a last time...
|
|
||
| using DomainView = | ||
| typename DataContainerType::domain_view_type::host_mirror_type; | ||
| specfem::datatype::DomainView<specfem::element::dimension_tag::dim2, |
| const int count = element_range.size(); | ||
|
|
||
| impl::write_coordinates(group, assembly.mesh, element_range); | ||
|
|
There was a problem hiding this comment.
I think this is still confusing. I suggest attenuation return names of properties that are attenuated. i.e. "kappa" & "mu". Any transformations that happen to the view kappa & mu should happen in this file so its easy to follow. I suggest that attenuation could return scaling factors that can be used here.
There was a problem hiding this comment.
The reason I set it up like this is that this one way of performing the scaling. For a different attenuation type that scaling may be different. So, if we make hardcode the scaling here, we would have to run scaling here every time.
I agree that this bit is confusing.
2 participants
Description
Here we go again!
Should address the original comments and is not modifying the
medium_writer.The properties IO tests have also been simplified compared to the original Attenuation IO implementation.
Issue Number
Closes #1810
Checklist
Please make sure to check developer documentation on specfem docs.