Datasets-SU2020-Naming-Conventions: Difference between revisions
Jump to navigation
Jump to search
No edit summary |
No edit summary |
||
Line 14: | Line 14: | ||
: to be used for all practical purposes in the software. Ok, so | : to be used for all practical purposes in the software. Ok, so | ||
dddddddd = | dddddddd = DDDDVSOB | ||
DDDD: 4-character-long name stub (it looks that 4 chars would allow us to name meaningfully all datasets we have so far dealt with | DDDD: 4-character-long name stub (it looks that 4 chars would allow us to name meaningfully all datasets we have so far dealt with | ||
V : dataset version, i.e. 0,1,2 ... (in principle, could be anything and be used for any purpose) | |||
S : stage index: some datasets require a multi-stage generation, to make a clean and visible cut, can assign <b>S=a</b> to reconstruction | S : stage index: some datasets require a multi-stage generation, to make a clean and visible cut, can assign <b>S=a</b> to reconstruction | ||
O : for stages with several output streams, just number them - 0,1,2,3,4... | O : for stages with several output streams, just number them - 0,1,2,3,4... | ||
B : beam intensity : 0 - signal only, 1 : 1 batch mode, 2 : 2 batch mode , the rest -reserved for future | B : beam intensity : 0 - signal only, 1 : 1 batch mode, 2 : 2 batch mode , the rest -reserved for future | ||
examples: | examples: |
Revision as of 19:34, 23 June 2020
(work in progress)
Mu2e files are required to have 6 field-long names, see https://mu2ewiki.fnal.gov/wiki/FileNames for more details, example:
cnf.mu2e.su2020.dddddddd.rrrrrr_ssssssss.art
- Let's use MC run number 1000 only for all datasets
- lets fix the third field to su2020 (lowcased) - that would allow to find handily for all su2020 datasets in SAM catalog
- 1+2 fix 5 fields out of 6.
- It makes sense to treat the remaining field as a 8 char long dataset ID. Having fixed length dataset ID is handy
- in many respects - we will need to propagate it through the whole analysis chain, from the MC files through the output
- of reconstruction to ntuples and histogram files.
- in addition to the dataset ID, a dataset has to have a long descriptive name, but it is the dsid which is supposed
- to be used for all practical purposes in the software. Ok, so
dddddddd = DDDDVSOB
DDDD: 4-character-long name stub (it looks that 4 chars would allow us to name meaningfully all datasets we have so far dealt with
V : dataset version, i.e. 0,1,2 ... (in principle, could be anything and be used for any purpose) S : stage index: some datasets require a multi-stage generation, to make a clean and visible cut, can assign S=a to reconstruction O : for stages with several output streams, just number them - 0,1,2,3,4... B : beam intensity : 0 - signal only, 1 : 1 batch mode, 2 : 2 batch mode , the rest -reserved for future examples:
|------------------------+---------------------------------------------------| | 4 char dataset ID stub | comments, examples | |------------------------+---------------------------------------------------| | beam | beam, trace everything | | bmuo | bmuo : beam, trace muons, but not e+,e-, gamma | | | bmuo1000 : s1, 0=first output stream - into TS | | | bmuo2000 : | | | bmuo3000 : | | bpim | beam pi- tracing | | mdio | DIO electrons, | | | examples: dioe01s5: leading order | | | dioe02s5: leading log | | mumi | mu-, example : mumi01s5 single particle gun muons | | mupl | mu+ | | elec | e- (i.e. single particle gun) | | posi | e+ | | pipl | pi+ | | pimi | pi- | | pbar | pbars example: pbar0110 (reconstructed) | | rpce | RPC external | | rpci | RPC internal | | rmce | RMC external | | rmci | RMC internal | | cosm | cosmics | | prot | ejected protons | | deut | ejected deutrons | | phot | ejected photons | |------------------------+---------------------------------------------------|