Clustering: as_pctree would fail if the PCTree root has 0 child

[HaiwangYu]

This seems to be fixed with this 80b80a3
But I don't understand if that is truely fixed and why so.

It can be reproduced by running SBND clustering with Ewerton's neutrino only simulation where one APA can have no activities.

Instruction on gpvm:
https://github.com/HaiwangYu/wcp-porting-img/blob/main/sbnd/README.md

command:

lar -n1 --nskip 1 -c wcls-img-clus.fcl /exp/sbnd/data/users/ebatista/wirecell-clustering/out/enuelastic/sigproc_20evts.root -o tmp.root

Ewerton's report:
Presentation2.pdf

Haiwang's notes:
clustering-2025-01-23_Ewerton-crash.pdf

Currently, I do not have an easy way to reproduce that in a standalone way.

[brettviren]

@HaiwangYu

Where do I find the pgrapher/experiment/sbnd/wcls-img-clus.jsonnet?

I do not find it in WCT's cfg/.

[HaiwangYu]

@brettviren Sorry I missed your comment earlier.

The jsonnet is here:
https://github.com/WireCell/wire-cell-toolkit/blob/ewerton-cfg/cfg/pgrapher/experiment/sbnd/wcls-img-clus.jsonnet
And is included in this PR:
#362
How about we just merge this PR?

[brettviren]

@HaiwangYu I think merging is fine. I guess something is wrong to make that empty entry but if simply skipping it lets the code run okay, we can figure out WHY it happens in the first place later. Let's leave this ticked open until we do the work to get that understanding.

[HaiwangYu]

I agree. Thanks!