jobsub: improved auto-naming of config files
This closes #57 (closed).
Previously, each line of the csv file corresponded to one job. The problem was when analysing one run multiple times, one had to fill multiple lines in the csv file like this:
RunNumber, MyParameter
100, 10ns
100, 12ns
100, 14ns
which was translated into config files like
run100_0.conf
run100_1.conf
run100_2.conf
where _0
was just an iterative number (and more problematically was overridden when re-running only a subset of the above runs for example).
This MR introduces the following new/improved syntax for the csv file:
-
"{10,12-14}"
translates to10, 12, 13, 14
in consecutive jobs for the same run number -
"{10ns, 20ns}"
tranlates to10ns, 20ns
in consecutive jobs for the same run number -
"string,with,comma"
translates tostring,with,comma
-
\
as escape character such that also"
can be part of the string
and the config files are then named like this (in case of example 2):
run100_parameter10ns.conf
run100_parameter20ns.conf
However, the usage of a range or set of parameters via {}
is only possible for on parameter in each run.
To Do:
-
run some more tests -
update README on how to use it
Edited by Jens Kroeger