spirit-code / spirit

Atomistic Spin Simulation Framework

http://spirit-code.github.io

MIT License

118 stars 52 forks source link

CI: add linting #618

Closed GPMueller closed 1 year ago

GPMueller commented 2 years ago

[x] add C++ format linting and clang-tidy reporting for the core library
[x] add C++ format linting and clang-tidy reporting for the ImGUI
[x] add Python format and quality linting for the core library wrappers and tests
[x] add Python format and quality linting for the ui-python folder

codecov[bot] commented 2 years ago

Codecov Report

Merging #618 (24c4cc3) into develop (f19764f) will increase coverage by 0.31%. The diff coverage is 54.94%.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## develop #618 +/- ## =========================================== + Coverage 48.16% 48.48% +0.31% =========================================== Files 74 75 +1 Lines 11968 12015 +47 =========================================== + Hits 5764 5825 +61 + Misses 6204 6190 -14 ```

coveralls commented 2 years ago

Coverage: 80.833% (+0.1%) from 80.705% when pulling 24c4cc3049af0e433c01300dcfa244b9391da1e7 on feature-ci-linting into f19764ff0653bdd62b7e54db656e23b873107c65 on develop.

GPMueller commented 1 year ago

I've managed to fix the CI, except:

Depondt and SIB solvers don't achieve the expected precision when using OpenMP (the Depondt solver can be fixed with c982a34c604aa763bb615123be6fb442738a089c)
test_anisotropy.cpp contains a test comparing energies from Energy and Gradient_and_Energy, which only succeeds when the precision is lowered to 1e-4

GPMueller commented 1 year ago

@MSallermann I believe this PR is now ready

MSallermann commented 1 year ago

I've managed to fix the CI, except:

* Depondt and SIB solvers don't achieve the expected precision when using OpenMP (the Depondt solver can be fixed with [](https://github.com/spirit-code/spirit/commit/c982a34c604aa763bb615123be6fb442738a089c)

* `test_anisotropy.cpp` contains a test comparing energies from `Energy` and `Gradient_and_Energy`, which only succeeds when the precision is lowered to `1e-4`

I have pushed another commit to the suzuki_trotter branch https://github.com/spirit-code/spirit/commit/04e6b8983fc8754c6b9986eb01309c594eb7f981. With this, and the commit you mentioned, I can set epsilon for the Larmor precession test to 1e-9 (as opposed to 1e-6 at the moment) and all solvers pass.
The problem with the last assert in test_anisotropy is that a float and a double are compared and the float truncates the last digits