feat: Schema for the oldstyle edm4hep Future Circular Collider simulation Samples

[prayagyadav]

Hi,
This is a schema for the Future Circular Collider Spring2021 pre-generated samples based on oldstyle EDM4HEP. This is a work in progress and I wanted to post it here for community inputs.
I have added most of the features:

• Added mixing classes most of the EDM4HEP classes
• Association between ReconstructedParticle and MCParticle
• Methods to get the daughters and parents to each MCParticle
• Special methods for Tracks and Clusters (Need input from the community)
• Tests and test root file for the FCCSchema

Further development will need inputs from more people.
Thanks

Links and References:

• Used p8_ee_ZH_ecm240 sample from here for testing (generated a smaller file with 100 events)
• EDM4HEP
• Structure of EDM4HEP files

Tagging my guides: @davidlange6 @gomber

[lgray]

Hi! Sorry for taking some time to get back to you on this.

This schema looks really well put together, thank you, and thanks for adding some basic tests.

I'll get through a review soon.

[lgray]

It's a bit less clean than yours but you may also want to have a look at #822. It's more a work in progress though.

[lgray]

What would be everyone's opinion on generalizing this to any edm4hep? It should be relatively easy, mostly you have to deal with missing branches as in NanoAOD.

[prayagyadav]

@lgray Thanks for taking the time to review. #822 was indeed, a starting point for the development of FCCSchema. Many pre-generated samples in FCC Events, especially the Spring2021 and Winter2023, are pre-edm4hep1(i.e. the oldstyle EDM4HEP I believe).

Yeah, creating a similar schema for the FCCSamples based on new style edm4hep (eg. file generated by @davidlange6) must be fairly easy.

Looks like #822 describes a general edm4hep schema, but the names of the branches assumed there, do not match with ones in FCC events. It would be really good if we could have a general edm4hep schema, and fcc schema could inherit from that, adding fcc specific functionality.

Nevertheless, the current FCCSchema( oldstyle edm4hep ) needs to be there, to facilitate interpretation of the older campaigns (Spring2021, Winter2023, etc.)

@davidlange6 What's your opinion on this?

[lgray]

Oh those were some initial comments. I have more that I want to go through, but it'll be monday or so.

[davidlange6]

What would be everyone's opinion on generalizing this to any edm4hep? It should be relatively easy, mostly you have to deal with missing branches as in NanoAOD.

right - its the eventual goal.. @prayagyadav and I thought to use this as an initial version to get feedback (and eventually users) on

[lgray]

@tmadlener FYI this is was I was talking about yesterday.

[prayagyadav]

Hi! @lgray , I figured out how to add global indexers and cross-references. I think I have fixed all of the issues that you had pointed out. Sorry for the unsquashed commits (I thought I did squash them, but evidently, it did not work). I think I will add some more tests later to check the validity of parents and daughters cross references. For now, I have this simple test in this notebook.
Please let me know what do you think about the changes. Thank you.

[lgray]

@prayagyadav Thanks for following this up despite lack of feedback. I was busy at a workshop. What you have implemented looks correct.

I'll wait for your additional tests and then we'll merge!

[lgray]

Closes #822

[prayagyadav]

Hi @lgray, I have added the tests for parents and daughters. I have also added one more sample file (needed to test the winter2023 campaign of FCC-Events).

[lgray]

@prayagyadav For your 4th checkbox, do you want to address that in a later PR, or are you waiting for some more prompt feedback?

[prayagyadav]

@prayagyadav For your 4th checkbox, do you want to address that in a later PR, or are you waiting for some more prompt feedback?

I can address that in a later PR when I figure out that part of the functionality.

[lgray]

@prayagyadav all looks good - auto-merge enabled on this latest CI run. Thanks a lot for your work!

We should plan with @davidlange6 and @tmadlener about extending this to new-style edm4hep.

I have made a tracking issue for the Tracks and Clusters functionality. We'll follow up in time.

[tmadlener]

For completeness, and because it might be useful for development: This workflow of EDM4hep creates a dummy EDM4hep file that contains all available datatypes and relations between them. That should give you all the branch names (and actual c++ datatypes if that matters here) for EDM4hep.

[prayagyadav]

@prayagyadav all looks good - auto-merge enabled on this latest CI run. Thanks a lot for your work!

We should plan with @davidlange6 and @tmadlener about extending this to new-style edm4hep.

I have made a tracking issue for the Tracks and Clusters functionality. We'll follow up in time.

Awesome!