[Rivet] agile and alpgen

[Gavin Hesketh]

Hi all,
I got around this problem with a fix to HepMC (in v2.6.4).

The input to agile should be a sample of N unweighted events from 
alpgen. You can now simply tell agile to generate N events. Some 
fraction of these will fail the matching, so the output of agile will be 
n (<N) real events, after which pythia/herwig will exit cleanly. Agile 
will then produce (N-n) empty events.
Then (thanks to the HepMC fix), as soon as rivet tries to read in an 
empty HepMC event, HepMC knows the run has finished and isses the end of 
run command, and everything exits cleanly. Rivet only sees the n 
successfully matched events.

So, there is a way to work around this. I suppose this might be a 
problem if agile is writing events to a pipe - if rivet exit and agile 
still wants to write more empty events, it may just sit in the 
background not doing anything until it gets killed. I didn't find this 
to be a problem though.

cheers,
	Gavin


On 13/04/11 00:57, Andy Buckley wrote:
> On 13/04/11 00:16, Anton Karneyeu wrote:
>> And last question for today - could you advice me a way to let
>> agile-runmc work till the last event in the input stream? This is
>> necessary for alpgen generator, when you don't know in advance number of
>> events which will come to hadronization step.
>
> I think that AlpGen is the only generator interface which has a concept
> of an input stream, so the AGILe framework really has no idea that such
> a thing is happening.
>
> I think that making this work could require some development, e.g.
> making use of the return value for the makeEvent() method to stop the
> event loop -- if you're using it, I think it might be best for you to
> try making such a patch, and if it works we'll integrate it for the next
> release.
>
> Andy
>