Sorry, am in US for some meeting so communication is difficult... let me have a look.

I think I know what is going on, Can you check - in MAUS output what is the event number and particle number? I think that I do global cuts by event number only (i.e. was this particle in bunch 1 the same as the particle in bunch 2?), but in MAUS I ID hits by event number = Spill and particle number = primary - which comes from retrofitting the MAUS data structure to an existing python code.

I can make a feature to add particle number to the global cut logic. It is a little complicated as I need to go into the C code... workaround is to assign manually (sorry) event numbers. But can you verify that this is what is going on?

Yep looks like event_number = Spill Number for the Virtual Hits (i.e. 1 in the example I posted).

Thanks

Hey Chris,

At the moment I'm looping over all hits in all planes & applying rather rudimentary cuts (w/o a transmission cut)...and it's all rather slow. Will the C code mod. speed things up also? Do you have an eta? I can then get stuck into checking step length, and [hopefully] present @ the next Analysis meeting.

Thanks

What are your cuts? Is the slow bit assigning event numbers, making the cuts or looking at the output? Or something else?

I simply loop over all hits in all planes & cut (v. basic, slow):
@
bunch_list = Bunch.new_list_from_read_builtin(filetype, filename)

for b in range( len(bunch_list) ):
print "Plane " + str(b)
for a in range( len(bunch_list[b]) ):
bunch = bunch_list[b]
bunch.cut({'y':-300.}, operator.le, global_cut=False)
bunch.cut({'y':300.}, operator.ge, global_cut=False)
bunch.cut({'x':-300.}, operator.le, global_cut=False)
bunch.cut({'x':300.}, operator.ge, global_cut=False)
bunch.cut({'pz':150.}, operator.le, global_cut=False)
@
Is there a simpler/faster way, and one that enables me to make a transmission cut?

is this too fiddly/time consuming, or still live? Thanks

Nope, sorry I was distracted onto other things. Will try to get it done this week.

So I implemented a native C routine for the inner loop of cut(...). Code and tests are committed to xboa launchpad as revision 89, and will go in the next release when it is ready.

bzr://xboa.bzr.sourceforge.net/bzrroot/xboa

Note that this is my development version - I think it is okay, failing tests to do with loading data from an obscure file format, but otherwise should be okay. I will just run some time trials on it also to see whether it is really any faster.

Thanks! Don't think I'll use it for anything other than 6D variables. I'll try it out next week.

See also #612