Fix check for MG event_norm
Description of bug
Currently for the event_norm
MG check logParser
just loops through the lines in log.generate
and it would throw a warning as soon as it encounters event_norm = sum
. In some cases event_norm = sum
is the default and is only changed later in the log.
Example from ATLMCPROD-8252:
01:24:20 Py:MadGraphUtils INFO Setting event_norm = sum
and further down
01:24:23 Py:MadGraphUtils INFO Setting event_norm = average
Changes introduced
Use
Py:MadGraphUtils INFO Setting event_norm
to extract how MG normalisation is set and read the last occurrence of the relevant dictionary entry.
Tests
!438 (merged)
Before fix with jO used inpython scripts/logParser.py -i 501xxx/501138/log.generate
-------------------------
Generator specific tests: MadGraph(v.2.8.1p3.atlas6)+Pythia8(v.244p3.rangefix)+EvtGen(v.1.7.0)
-------------------------
INFO: Pythia 8 checks: Passed
WARNING: The use of event_norm=sum will almost always result in the sample having the wrong total cross section -- please double check that event_norm=average is set in the param_card.dat.
INFO: MadGraph checks: Passed
After fix
python scripts/logParser.py -i 501xxx/501138/log.generate
-------------------------
Generator specific tests: MadGraph(v.2.8.1p3.atlas6)+Pythia8(v.244p3.rangefix)+EvtGen(v.1.7.0)
-------------------------
INFO: Pythia 8 checks: Passed
INFO: MadGraph checks: Passed
@hmildner @zmarshal it is a bit "annoying" that MadGraphUtils
prints out stuff twice.
I didn't think this would happen so it is quite likely that other checks in logParser
might be broken.
@zmarshal @hmildner Is it possible to only do the printout at the end after all parameters are set?
We can also change the checks in logParser however it would seem cleaner to only have 1 printout.
Issues resolved
Closes #129 (closed)