The code crash when using SMEFT@NLO to produce Higgs Events

[Gangcheng Lu]

Hi,

When I used the SMEFT@NLO to perform Higgs production, the code crashed like:
Command "import /publicfs/atlas/atlasnew/higgs/hgg/lugc/MG5_aMC_2021/MG5_aMC_v3_1_0/proc_card_ggF_linear_quad_NLO.dat" interrupted in sub-command:
"output SMEFT_ggF_linear_quad_CHG_NLO" with error:
KeyError : 1
Please report this bug on https://bugs.launchpad.net/mg5amcnlo
More information is found in 'MG5_debug'.
Please attach this file to your report.

The MG5 version I used is latest v3.1.0, and SMEFT@NLO version is v1.0.2. I have attached the Bug Report below the message

My syntax is very simple:
import model SMEFTatNLO-NLO
generate p p > h NP=2 QED=1 QCD=2 [QCD]

It seems ok when I execute the second syntax in the interactive shell of mg5_aMC, but errors occurred if I want to output the directory.

When I used the command displayed in official documentation:https://feynrules.irmp.ucl.ac.be/wiki/SMEFTatNLO
g g > H QED=1 QCD=2 NP=2 [QCD] (replace the "p p" in old syntax with "g g")
The errors still existed. Do you have any ideas on that? thanks for the answers in advance!

[Olivier Mattelaer (olivier-mattelaer)]

Hi,

I assign Eleni, since she does have a patch for that (looks like very complex issue).

Cheers,

Olivier

[marco zaro (marco-zaro)]

Hi,
I tried your generate command, and got this:
MG5_aMC>generate g g > h NP=2 QED=1 QCD=2 [QCD]
Interpreting 'QCD=2' as 'QCD<=2'
Interpreting 'QED=1' as 'QED<=1'
Interpreting 'NP=2' as 'NP<=2'
Order QCD is not constrained as squared_orders. Using: QCD^2=4
Order QED is not constrained as squared_orders. Using: QED^2=2
Order NP is not constrained as squared_orders. Using: NP^2=4
---------------------------------------------------------------------------
 No Born diagrams found. Now switching to the loop-induced mode.
 Please also cite ref. 'arXiv:1507.00020' when using results from this mode.
---------------------------------------------------------------------------
Interpreting 'QCD=2' as 'QCD<=2'
Interpreting 'QED=1' as 'QED<=1'
Interpreting 'NP=2' as 'NP<=2'
1 processes with 12 diagrams generated in 0.137 s
Total: 1 processes with 12 diagrams

then, at the output, got
…
INFO: Generating Feynman diagrams for Process: g g > h NP<=2 QCD<=2 QED<=1 [ noborn = QCD ]
Command "output " interrupted in sub-command:
"output" with error:
KeyError : 1
Please report this bug on https://bugs.launchpad.net/mg5amcnlo
More information is found in 'MG5_debug'.
Please attach this file to your report.

    File "/Users/marcozaro/Physics/MadGraph/3.2.0/madgraph/interface/master_interface.py", line 313, in export
      return self.cmd.export(self, *args, **opts)
    File "/Users/marcozaro/Physics/MadGraph/3.2.0/madgraph/interface/madgraph_interface.py", line 8278, in export
      calls += self._curr_exporter.export_processes(self._curr_matrix_elements,
    File "/Users/marcozaro/Physics/MadGraph/3.2.0/madgraph/iolibs/export_v4.py", line 198, in export_processes
      calls = calls + self.generate_subprocess_directory(\
    File "/Users/marcozaro/Physics/MadGraph/3.2.0/madgraph/loop/loop_exporters.py", line 3343, in generate_subprocess_directory
      export_v4.ProcessExporterFortranMEGroup.generate_subprocess_directory(
    File "/Users/marcozaro/Physics/MadGraph/3.2.0/madgraph/iolibs/export_v4.py", line 5734, in generate_subprocess_directory
      self.write_coloramps_file(writers.FortranWriter(filename),
    File "/Users/marcozaro/Physics/MadGraph/3.2.0/madgraph/iolibs/export_v4.py", line 5982, in write_coloramps_file
      lines.extend(self.get_icolamp_lines(subproc_to_confdiag[subproc],
    File "/Users/marcozaro/Physics/MadGraph/3.2.0/madgraph/iolibs/export_v4.py", line 1160, in get_icolamp_lines
      bool_list = [(i + 1 in diag_jamp[num_diag]) for i in range(colamps)]
    File "/Users/marcozaro/Physics/MadGraph/3.2.0/madgraph/iolibs/export_v4.py", line 1160, in <listcomp>
      bool_list = [(i + 1 in diag_jamp[num_diag]) for i in range(colamps)]
  KeyError: 1

Are you sure that you want to generate a loop-induced process?

Marco

> On 19 Apr 2021, at 15:05, Gangcheng Lu <email address hidden> wrote:
>
> Public bug reported:
>
> Hi,
>
> When I used the SMEFT@NLO to perform Higgs production, the code crashed like:
> Command "import /publicfs/atlas/atlasnew/higgs/hgg/lugc/MG5_aMC_2021/MG5_aMC_v3_1_0/proc_card_ggF_linear_quad_NLO.dat" interrupted in sub-command:
> "output SMEFT_ggF_linear_quad_CHG_NLO" with error:
> KeyError : 1
> Pleas...

[Gangcheng Lu (allaturca)]

Hi Marco,

Thanks for the check, that is actually what I encountered. Sure, I want to generate a loop-induced process and study the impacts with NLO contributions included. Since the syntax is just from the SMEFT@NLO official documentation, I have no idea why it doesn't work.

Best regards,
Gangcheng

[Olivier Mattelaer (olivier-mattelaer)]

Hi Gangcheng,

This is a well know issue that Valentin fails to solve since a while.
He provides a patch to Eleni, but likely not general enough to be included in the official version.

> study the impacts with NLO contributions included

We are not able to compute NLO correction on a loop-induced process (the most obvious reason being that we are not able to compute the associated two loop contribution).
All loop-induced process will be at LO accuracy.

Cheers,

Olivier

[Gangcheng Lu (allaturca)]

Hi Oliver,

Thanks for the clarification! Could you please share the patch to me, so that I can test whether it can solve my problem?

>We are not able to compute NLO correction on a loop-induced process (the most obvious reason being that we are not able to compute the associated two loop contribution).
All loop-induced process will be at LO accuracy.

I think maybe the characteristics of SMEFT@NLO is calculating loop diagram at NLO instead of the traditional LO calculation using leading-order perturbation?

Best regards,
Gangcheng

[Olivier Mattelaer (olivier-mattelaer)]

I do not have the patch. Eleni does.

>
> I think maybe the characteristics of SMEFT@NLO is calculating loop
> diagram at NLO instead of the traditional LO calculation using leading-
> order perturbation?

That's not possible to do neither the information within SMEFT@NLO or within mG5aMC framework.

Cheers,

Olivier

> On 21 Apr 2021, at 05:32, Gangcheng Lu <email address hidden> wrote:
>
> Hi Oliver,
>
> Thanks for the clarification! Could you please share the patch to me, so
> that I can test whether it can solve my problem?
>
>> We are not able to compute NLO correction on a loop-induced process (the most obvious reason being that we are not able to compute the associated two loop contribution).
> All loop-induced process will be at LO accuracy.
>
> I think maybe the characteristics of SMEFT@NLO is calculating loop
> diagram at NLO instead of the traditional LO calculation using leading-
> order perturbation?
>
> Best regards,
> Gangcheng
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1925029
>
> Title:
> The code crash when using SMEFT@NLO to produce Higgs Events
>
> Status in MadGraph5_aMC@NLO:
> Confirmed
>
> Bug description:
> Hi,
>
> When I used the SMEFT@NLO to perform Higgs production, the code crashed like:
> Command "import /publicfs/atlas/atlasnew/higgs/hgg/lugc/MG5_aMC_2021/MG5_aMC_v3_1_0/proc_card_ggF_linear_quad_NLO.dat" interrupted in sub-command:
> "output SMEFT_ggF_linear_quad_CHG_NLO" with error:
> KeyError : 1
> Please report this bug on https://bugs.launchpad.net/mg5amcnlo
> More information is found in 'MG5_debug'.
> Please attach this file to your report.
>
> The MG5 version I used is latest v3.1.0, and SMEFT@NLO version is
> v1.0.2. I have attached the Bug Report below the message
>
> My syntax is very simple:
> import model SMEFTatNLO-NLO
> generate p p > h NP=2 QED=1 QCD=2 [QCD]
>
> It seems ok when I execute the second syntax in the interactive shell
> of mg5_aMC, but errors occurred if I want to output the directory.
>
> When I used the command displayed in official documentation:https://feynrules.irmp.ucl.ac.be/wiki/SMEFTatNLO
> g g > H QED=1 QCD=2 NP=2 [QCD] (replace the "p p" in old syntax with "g g")
> The errors still existed. Do you have any ideas on that? thanks for the answers in advance!
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/mg5amcnlo/+bug/1925029/+subscriptions

[Eleni Vryonidou (evryonidou)]

You can substitute

diag_jamp = {}

with

diag_jamp = dict((i,[]) for i in mapconfigs)

around line 974 of export_v4.py

[Tae Hyoun Park (taehyounpark)]

Hello,

I've encountered the exact same issue and fortunately came across this post. After looking into this a bit more, I found that:

(1) "Force" SM-only:
  generate p p > h [noborn=QCD] QED=1 NP=2 NP^2==0

(2) Linear-only:
  generate p p > h [noborn=QCD] QED=1 NP=2 NP^2==2

(3) BSM-only:
   generate p p > h [noborn=QCD] QED=1 NP=2 NP^2==4

(4) BSM-only for ggZH:
  generate g g > h l+ l- [noborn=QCD] NP=2 NP^2==4

Interestingly, (1) and (2) runs fine. (3) produces the exact same error, whereas (4) with a "slightly different" process also runs fine.

I'm a bit puzzled by which specific scenarios under which this issue is being encountered by mg5aMC+SMEFT@NLO, and would greatly appreciate if the experts could shed some insight into this. Also, any feedback on which processes/syntax would yield obtainable & valid results (i.e. LO EFT correction for loop-induced processes) will be very helpful.

For reference, am using MG5_aMC@NLO 2.9.9 (inside ATLAS experiment's event generation environment) + SMEFTatNLO 1.0.3.