- removed comment
need tensorparity=-1 for WeylScal4::curvIi_group and curvJi_group
The WeylScal4 gridfunctions curvIi and curvJi are pesudoscalars with the same parity properties as Psi2r and Psi2i. However, they're currently assigned no tensorparity attribute, so they're treated as scalars.
As a result, for a z-aligned Kerr BH (EinsteinInitialData/IDAnalyticBH) evolved with reflection symmetry across the x-y plane, an interpolation of curvIi to a coordinate sphere will yield different values for the z<0 points than the same data on a full grid.
Can we set "tensorparity=-1" for both these groups, to fix this issue?
Keyword: weylscal4,
Keyword: parity
Comments (4)
-
-
- changed status to open
- removed comment
I have checked the parities for I and J, and agree with what "physicsbeany" suggests: i.e. curvIi and curvJi should be -1. curvIr and curvJr remain as +1. I have committed the corresponding change (https://bitbucket.org/einsteintoolkit/einsteinanalysis/commits/dbf613b8dd7c34c685e9ef9091c7608f146b9db4) and regenerated the code (https://bitbucket.org/einsteintoolkit/einsteinanalysis/commits/6eef88eef2e6dfb51d61e5b57e2f95ac814928ad).
There are also invariants //J1–J4 "of the Narlikar and Karmarkar basis as defined in B5-B8 of arXiv:0704.1756. Computed from Weyl tensor expressions using xTensor."// which I think Barry Wardell added. Barry, do you know what the parities of these are supposed to be? In my commit, I added a TODO for this. I have computed the parities, and get +1 for all of them. Do you agree? If so, I will add this explicitly in the code to eliminate the warning about assuming +1.
-
- changed status to resolved
- removed comment
-
- changed status to closed
- edited description
- Log in to comment
Sure, that should work. Do you want to open a pull request on Bitbucket?