Add openPMD-viewer test #4747
Add openPMD-viewer test #4747
Conversation
psychocoderHPC
added
component: core
|
Please reformat your python files to be |
max-lehman14
force-pushed
the
add_openPMDviewer_test
branch
from
846a13b
to
7920360
Compare
|
@PrometheusPi I am using openpmd_viewer 1.7.0 |
|
Please add a |
|
I tested the new test with ./picongpu/share/picongpu/tests/openPMD-viewer-test/bin/ci.sh picongpu/share/picongpu/tests/openPMD-viewer-test/ ./run01on hemera and it works fine. |
There was a problem hiding this comment.
There are a few (minor) code changes - see comments.
I tested the code to fail by changing the PIConGPU param files to disagree with the (python) parameter dictionary against it was compared.
That lead to the intended errors when:
- changing the background electric field
- changing the gamma from
1.021to2.021 - not adding ion species
When rebasing against the mainline dev, I assume one gets compile time errors.
(I currently could not test that due to high work load on hemera's v100)
I will test it myself asap, but_
--> Please rebase against the current dev
share/picongpu/tests/openPMD-viewer-test/lib/python/test/param-test.py
Outdated
Show resolved
Hide resolved
| eps_s = 1 | ||
| eps_l = np.abs( | ||
| field_param[f"{density}"] * 1e-1 |
There was a problem hiding this comment.
see comment before: please define this in your parameters dictionary (or a separate epsilon dictionary)
There was a problem hiding this comment.
what is the meaning of eps_s and eps_l please add a comment
There was a problem hiding this comment.
field_param[f"{density}"] * 1e-1
Why do we accept an error of 10%? Is this due to the particle motion or can this be reduced?
| ] | ||
| for item in array | ||
| ] | ||
| eps = 0.1 |
There was a problem hiding this comment.
see comment above on epsilons
|
|
||
| bool_pass = check_params(2) # false if failure | ||
| print("Test result: ", bool_pass) | ||
| sys.exit(not bool_pass) |
There was a problem hiding this comment.
please add a comment on why we needed to negate the return value - otherwise this is really confusing
|
|
||
| This test reads out openPMD-viewer data and checks if it matches the input values to see if users can rely on the output. | ||
|
|
||
| To run this test, one has to execute ci.sh with the location of the input and output directory. |
There was a problem hiding this comment.
Could you please add a codeblock to show that?
|
I rebased against the current |
max-lehman14
force-pushed
the
add_openPMDviewer_test
branch
from
ca16c00
to
8252eec
Compare
|
@max-lehman14 There are still compile errors in your code ee.g.: |
steindev
added
the
changelog
max-lehman14
force-pushed
the
add_openPMDviewer_test
branch
from
67a6e1d
to
2f732ba
Compare
|
@max-lehman14 the CI fails due to an undocumented change in behaviour in recent typeguard versions first noted in issue #4853 and fixed by PR #4861, please rebase to the current dev |
max-lehman14
force-pushed
the
add_openPMDviewer_test
branch
from
696477a
to
811fc38
Compare
There was a problem hiding this comment.
Looks very good @max-lehman14 just a few minor issues left.
share/picongpu/tests/openPMD-viewer-test/lib/python/test/param-test.py
Outdated
Show resolved
Hide resolved
|
@max-lehman14 I think your code base is too old and thus triggers a CI error that your actual code does not cause. |
max-lehman14
force-pushed
the
add_openPMDviewer_test
branch
from
8802a81
to
fc2f724
Compare
|
@max-lehman14 perfect - could you please squash all your commits to one. |
max-lehman14
force-pushed
the
add_openPMDviewer_test
branch
from
fc2f724
to
2c99a09
Compare
PrometheusPi
merged commit b23d490
into
ComputationalRadiationPhysics:dev
A simple example is used and the input parameters are compared to the readout of the openPMD-viewer. This test involves the electromagnetic fields, electron and ion charge density, particle positions and momenta.