Fix cone energy self-counting in IsolatedLeptonFinderProcessor

[yyzhang666]

BEGINRELEASENOTES

• Fixes several related issues caused by comparing copied reconstructed particles with originals by pointer identity
• Avoids self-counting in isolation cone energy
• Ensures jet-based isolation uses original PFOs
• Completes copying of tracks and clusters when duplicating PFOs

ENDRELEASENOTES

[tmadlener]

Thanks a lot for putting this together. A few comments / requests.

• This looks like it includes the changes from Bugfixes in SLDCorrection #143. Is that correct?
• There are quite a few merge conflicts in this PR. It might be that your work. This is almost certainly because it starts off quite a bit back from the master branch. This should probably be fixable by formatting the file accordingly, e.g. via clang-format -i -style=file Analysis/*/*/*.{cc,h}
• Please keep the original block from the template and fill the release notes in between them, as we use these tags to generate release notes for all PRs automatically (i.e. the BEGINRELEASENOTES ... ENDRELEASENOTES).

[yyzhang666]

Apology that I'm new to all git stuff.

* This looks like it includes the changes from [Bugfixes in SLDCorrection #143](https://github.com/iLCSoft/MarlinReco/pull/143). Is that correct?

I work from Bryan's repository so it may contains this #143 . I have rebased it to the current version officially here.

* There are quite a few merge conflicts in this PR. It might be that your work. This is almost certainly because it starts off quite a bit back from the master branch. This should probably be fixable by formatting the file accordingly, e.g. via `clang-format -i -style=file Analysis/*/*/*.{cc,h}`

Now I only formatted the IsolatedLeptonFinder, seems still some conflicts, I manually resolved them in Github here. But now sure if it is good, should I format all repository?

* Please keep the original block from the template and fill the release notes in between them, as we use these tags to generate release notes for all PRs automatically (i.e. the `BEGINRELEASENOTES ... ENDRELEASENOTES`).

Yes, I edit the first comment block and hope this is correct?

[tmadlener]

It looks like you have successfully picked only your changes here. There still seems to be quite a bit of additional formatting going on. In which (software) environment / setup are you running clang-format? In the end this doesn't matter too much for now, we can always fix it in the end, when all the other things are done.

I have a few smaller comments / requests below.

On the bigger picture it's not entirely clear to me why this processor creates copies of all PFOs. It also looks to me like they are (at least partially) leaked, but that seems to be pre-existing behavior. Did you by chance run into any memory issues when running this processor on larger samples?

[yyzhang666]

It looks like you have successfully picked only your changes here. There still seems to be quite a bit of additional formatting going on. In which (software) environment / setup are you running clang-format? In the end this doesn't matter too much for now, we can always fix it in the end, when all the other things are done.

It's what the ZHH repository are using: /cvmfs/sw.hsf.org/key4hep/releases/2024-10-03/x86_64-almalinux9-gcc14.2.0-opt/llvm/17.0.6-7m4ago/bin/clang-format

On the bigger picture it's not entirely clear to me why this processor creates copies of all PFOs. It also looks to me like they are (at least partially) leaked, but that seems to be pre-existing behavior. Did you by chance run into any memory issues when running this processor on larger samples?

I only run 12500 events once at most so far, looks normal. They basically use an additional collection to dress leptons and output a collection without isolated leptons.

[tmadlener]

It's what the ZHH repository are using: /cvmfs/sw.hsf.org/key4hep/releases/2024-10-03/x86_64-almalinux9-gcc14.2.0-opt/llvm/17.0.6-7m4ago/bin/clang-format

Ah, OK. Yeah that then explains it. The version that is used by the pre-commit workflow is 20, so there are some differences. In this case, I will fix the formatting at the end, once we have dealt with the rest. :)

[tmadlener]

I will push a commit with fixed formatting (which should also make the diff slightly smaller). I will have another look after that, but I think this is in pretty good shape now.

[tmadlener]

I would move this comment further up. Alternatively, remove it entirely as it is pretty obvious from the code what we are doing, but the why we are doing it is missing. In this case the why is probably simply that the _rpJetMap stores the original PFOs, so also that is not extremely interesting, resp. can also be derived from the code in a rather straight forward manner.

[tmadlener]

Suggested change

// FIXED

Looks like another leftover from development.

[yyzhang666]

Yes, those comments was written earlier and forgot to arrange them