Possible issue with FatJet JME module

[alpakpinar]

Hello, I had an issue running the latest nanoAOD-tools, I tried to summarize the problem I had in PR #253.

I believe the issue I had was related to the following: In the central version the JMS,JER,JMR branches for fat jets are initialized both in data and MC in 1. But then these are only filled for MC in 2. I believe this causes an issue with the following collectionMerger module (in our specific case): It sees that those branches are defined and looks for them, but cannot find anything and raises an error. I shared the full traceback earlier in #253.

I tried running by making two small differences, and for each of them I didn't get a problem:

1. Creating and filling these branches (with 1's) for data
2. Stop creating these branches for data

Did anyone else encountered a similar issue? For reference, the script I'm running is in 3. I encounter the problem for the collectionMerger module defined at L109. Please let me know if I can provide additional information, and thanks a lot for investigating this!

[gouskos]

Hi @alpakpinar Thanks for reporting this issue. I am fine with @alefisico suggestion to not try top merge PRs now given the upcoming changes from the JME land, but I have 2 points I would like to add here [not necessarily related to your PR, but more for JME developers]:
(a) I do not see any real reason for creating these branches for data.
(b) the JMS, JMR are only validate for specific selection on a specific tagger [i.e. tau21 IIRC]. I think in the next version of the JME tool we should add the flexibility to include these corrections for other taggers/working points

[alpakpinar]

Hi @gouskos , sure, thank you for following up. To make approaching the problem possibly easier, I prepared a more minimal script which reproduces the problem (thanks for suggesting this @AndreasAlbert). In the script in 1, I just apply the fatjet JME module and the collectionMerger module following that, and I encounter the same problem.