Open tomsail opened 4 months ago
I might have broken SCHISM tests with mpi4py
.
I think it might be relevant to create a fourth lock:
SCHISM_LOCK: locks/conda-${{ matrix.os }}-schism_${{ matrix.mpi }}.lock
DELFT3D_LOCK: locks/conda-${{ matrix.os }}-delft3d_${{ matrix.mpi }}.lock
PYPOSEIDON_LOCK: locks/conda-${{ matrix.os }}-binary-p${{ matrix.python }}.lock
PYPOS_TEL_LOCK: locks/conda-${{ matrix.os }}-binary-p${{ matrix.python }}_tel.lock
So I don't mess up with the other CIs
I think that:
mpi4py
in pyproject.toml pip install mpi4py
in the CI
is very bad .. but I couldn't find another way to make CI workI updated all the lock files. The CI seems to be happy. The telemac are marked as XPASS though. @tomsail did you expect them to fail for some reason?
I updated all the lock files. The CI seems to be happy. The telemac are marked as XPASS though. @tomsail did you expect them to fail for some reason?
No sorry this is an error. I had put it at the beginning and forgot to take it off.
Since the 4 tests (3 TELEMAC2D and 1 TOMAWAC) pass without problems I can remove it now.
@tomsail if you have any updates on your local branch you can push them. We will need to rebase and we will need to regenerate lock files etc.
@pmav99 let me know if you want me to rebase or if you want to check it too
I still need to regroup the telemac options in parameters, and standardize them with the ones used for SCHISM.
If you compare the specific parameters for SCHISM and TELEMAC in the same dictionary:
model = {
...
"parameters": {
"dt": 400,
"rnday": 30,
"hotout": 0,
"ihot": 0,
"nspool": 9,
"ihfskip": 36,
"hotout_write": 108,
}, # FOR SCHISM
"tstep": 400, # for telemac
"meteo_input360": True, # if meteo files longitudes go from from 0 to 360
}
For example here tstep
should be dt
and be placed in parameters.
I'll correct this in the next days
fail with python 3.12 is normal. There has been some update regarding this issue actually: https://gitlab.pam-retd.fr/otm/telemac-mascaret/-/issues/946
let me battle with CI before starting reviewing it. There is no rush anyway