feat: adding groomed jet radius and momentum balance variables for softdrop groomed jet

[kawaho]

resolves #303

[matthewfeickert]

@kawaho In the future please do not make PRs from your main branch and always use a feature branch. c.f. https://hynek.me/articles/pull-requests-branch/ Using your main branch for PRs makes it impossible to rebase your branch against the upstream without closing the PR.

[rkansal47]

Looks pretty good @kawaho!

Unfortunately there is a runtime error with the function now https://github.com/scikit-hep/fastjet/actions/runs/10819914062/job/30018826074?pr=312#step:7:135:

self = <fastjet._singleevent._classsingleevent object at 0x116f66a90>, njets = 1
beta = 0.0, symmetry_cut = 0.1, symmetry_measure = 'scalar_z', R0 = 0.8
recursion_choice = 'larger_pt', mu_cut = inf

    def exclusive_jets_softdrop_grooming(
        self,
        njets=1,
        beta=0.0,
        symmetry_cut=0.1,
        symmetry_measure="scalar_z",
        R0=0.8,
        recursion_choice="larger_pt",
        # subtractor = 0,
        mu_cut=float("inf"),
    ):
        # if njets <= 0:
        #    raise ValueError("Njets cannot be <= 0")
>       np_results = self._results.to_numpy_softdrop_grooming(
            njets,
            beta,
            symmetry_cut,
            symmetry_measure,
            R0,
            recursion_choice,  # subtractor,
            mu_cut,
        )
E       RuntimeError: std::bad_cast

../../../hostedtoolcache/Python/3.8.18/x64/lib/python3.8/site-packages/fastjet/_singleevent.py:215: RuntimeError

Perhaps the types for the new variables is incorrect?

Also, can you update the test_exclusive_jets_softdrop_grooming function as well please.

[kawaho]

@rkansal47 I think the types for the new variables are fine but the error comes from the use of dynamic_cast in https://github.com/alisw/fastjet/blob/master/fastjet/include/fastjet/PseudoJet.hh#L1126 . It is also strange because the error only happens for macOS, not sure how to resolve it yet....

[kawaho]

@rkansal47 I think the types for the new variables are fine but the error comes from the use of dynamic_cast in https://github.com/alisw/fastjet/blob/master/fastjet/include/fastjet/PseudoJet.hh#L1126

I have done more tests, separating out the soft.structure_of<fastjet::contrib::SoftDrop>() call, and can confirm the RuntimeError: std::bad_cast is from the dynamic_cast. But not sure how to fix yet.

[matthewfeickert]

@kawaho I've updated this PR to bring in the state of the upstream HEAD, so if you pull your branch will now have the changes from PR #311 in it which should make local development easier. To have a fully clean repository you'll still need to do the git clean steps of Issue #279, but you should be able to do new builds by just running

python -m pip install --upgrade .

without having to run the cleaning steps from Issue #279. If you have questions just ask here. 👍

[kawaho]

@matthewfeickert Thanks for the PR, it works great.

[kawaho]

@matthewfeickert and @rkansal47 I would like to hear suggestions on solving this failed test (https://github.com/scikit-hep/fastjet/actions/runs/10819914062/job/30018826074?pr=312#step:7:135). The RuntimeError std::bad_cast is coming from this specific line calling the dynamic casting in (https://github.com/alisw/fastjet/blob/master/fastjet/include/fastjet/PseudoJet.hh#L1126). The error thrown is happening on MacOS only. I can replicate this error on my own Mac but calling the same function in C++ only script with the same fastjet-contrib version did not give such an error. So I suspect this is related to pybind or how we compile the code. I am unfamiliar with both and would like to hear suggestions for things to try from any experts on this project.

[matthewfeickert]

Probably want to ask (in addition to @rkansal47) other people that have touched the C++: @jpivarski @jmduarte @chrispap95 @lgray @aryan26roy