hepmc not read by Delphes

[Amin Abou Ibrahim]

Dear Delphes team,

I installed Delphes 3.4.1 from MG5 2.6.0 prompt along with pythia8. With the systematics turned on, the resulting hepmc file from pythia8 cannot be read by Delphes:

** Reading /GlobalHome/aibrahim/MG5_aMC_v2_6_0/bin/SUSY_FULL/Events/run_01/tag_1_pythia8_events.hepmc
** [---------------------------------------------------------------- (0.00%)
** ERROR: invalid event format
** [################################################################] (100.00%)
** Exiting...

I have seen this problem mentioned in a previous ticket and it says it was fixed but I'm not sure why I'm getting this.

Thank you,
Amin

[Pavel Demin]

Could you please attach to this ticket the hepmc file?

[Amin Abou Ibrahim]

Hi Pavel,

The file is around 1.3 GB in size. Here's a google drive link to it:
​https://drive.google.com/open?id=1iQrlrGgGT6stu_jcVL6kGtj_rGs3tj-N

Thank you,
Amin

[Pavel Demin]

Hi Amin,

Thank you for the file.

The problem is with the following line in the file:

E 0 -1 -1.0000000000000000e+00 -1.0000000000000000e+00 -1.0000000000000000e+00 0 0 94 1 2 0 7 3.1671032000000003e-05 3.1671032000000003e-05 3.1671032000000003e-05 3.0000000000000000e+01 4.1375492312388307e+00 1.0000000000000000e+10

The format of this line seems to be invalid. The number of weights is set to 7 but there are only 6 weights. The weights are the last 6 floating point values at the end of the line.

I'd say that something should be fixed in the program that generates this file.

Best regards,

Pavel

[Amin Abou Ibrahim]

Hi Pavel,

Thanks for looking into it. I searched online for some help or fixes but I couldn't find.
When you say that something needs to be fixed in the program that generated the file, do you mean a bug might have caused this or something wrong due to user input?

Best,
Amin

[Pavel Demin]

When you say that something needs to be fixed in the program that generated the file, do you mean a bug might have caused this or something wrong due to user input?

I don't know how exactly the HepMC file is produced. Is it produced by Pythia 8 or by some other code in MadGraph 5? Anyway, I'd suggest to notify the MadGraph developers about this problem.

[Pavel Demin]

I've just done a quick test with MG5 2.6.0.

Here are the commands that I typed in the MG5 command line:

install pythia8
generate p p > e+ e-
add process p p > e+ e- j
output
launch
1
0

The resulting HepMC file can be read by Delphes without any problem.

Here is how the lines with weights look like:

E 0 -1 -1.0000000000000000e+00 -1.0000000000000000e+00 -1.0000000000000000e+00 0 0 1121 1 2 0 6 1.0185641000000001e-01 1.0185641000000001e-01 1.0185641000000001e-01 3.0000000000000000e+01 3.0000000000000000e+01 1.0000000000000000e+11
N 6 "Weight_MERGING=45.000" "Weight_MERGING=67.500" "Weight_MERGING=90.000" "scomp=3.0000000000e+01" "smax=3.0000000000e+01" "smin=1.0000000000e+11"

[Amin Abou Ibrahim]

Hi Pavel,

I had done a similar thing but for a BSM process. I have opened a ticket with MG5 but I'm still waiting for an answer.
It could be that there is an intricacy in the run_card that I am overlooking when I switch systematics on.

Thank you for investigating the issue.

All the best,
Amin

[Jake]

Hi Pavel and Amin,

I'm having the same "invalid event format" issue with Delphes 3.3.3. I have checked that I have as many weights as is specified in the header so this doesn't seem to be the issue. I've tried with more recent versions of Delphes, and this issue goes away, but I have problems with the number of photons reconstructed in newer versions so I would prefer to implement a solution within Delphes 3.3.3.

Amin, could you post the link to the MG5 ticket here please?

Thanks,
Jake

[Amin Abou Ibrahim]

Hi Jake,

My issue was created when I switch on systematics while performing matching/merging with pythia8. The resulting hepmc file does not have vetoed events but rather some modification of the weights is done to account for the matching/merging process. It seems that the problem arises here. I have opened a ticket with MG5 but I got no answer:

​https://answers.launchpad.net/mg5amcnlo/+question/660911

I simply switched off systematics for now in MG5 run_card to circumvent the problem.

Best,
Amin