JP tagger recalculation #77
Conversation
There was a problem hiding this comment.
Hi @colizz thank you for the PR on the JP recalculation!
Sorry I take a bit longer trying to figure out the possibility to get rid of the array builder if possible.
I only have some minor comments regarding to the restructure of the files and directory.
Please let me know if you have any suggestions! (I could also made these changes and merge it)
Thank you !
There was a problem hiding this comment.
If I understand correctly, the JP calibration change with different campaign. Could you please create a directory with Summer22EERun3 under data/JPCalib/
There was a problem hiding this comment.
Same comments w.r.t. other files
There was a problem hiding this comment.
I would suggest to put the whole macro of JPCalibHandler to utils.correction (good to be global variables also checked by commissioninig team)
|
|
||
| # print(f'Using JPCalib file {filename}') | ||
|
|
||
| templates = uproot.open(f"src/BTVNanoCommissioning/data/JPCalib/{filename}") |
There was a problem hiding this comment.
filename includes campaign info
|
Hi, I made some changes as recommended. I also managed to use the columnar operations (removing the array builder) in the JP calculation. |
Adding the calculation of jet probability (JP) and jet B probability (JBP).
This is the exact implementation in cmssw, but now written in coffea code.
For each AK4 jet, we take the matched tracks from the PFCands collection.
The IPsig templates as "JPCalibration" file are required to derive the track probability. These track probabilities are then used to calculate jet probability.