problem with continuation lines in input files (probably in MadSpin code)

[Kilian Rosbach]

I've come across an unusual error message while running a simple MadGraph+MadSpin event generation (I believe the details are unimportant, but I can provide them if requested): MadSpin fails, complaining about an unknown particle "dl~dr~" (notice the missing space between what should actually be two SUSY particles).

The exact message is: "InvalidCmd : No particle dl~dr~ in model"

Studying the input files, I found the following continued (with \ at the end) line:
"define susystrong = go ul ur dl dr cl cr sl sr t1 t2 b1 b2 ul~ ur~ dl~\
 dr~ cl~ cr~ sl~ sr~ t1~ t2~ b1~ b2~"
in the file PROC_mssm_0/Cards/proc_card_mg5.dat

It appears that the space following the backslash-newline combination is lost while reading this file. The problem disappears if I manually replace the file by one that doesn't use continued lines instead, but this is very difficult to do when running on the grid (the file is automatically generated by a script outside of my control).

Is this easy to fix on your end?

Thanks and best regards!

[Olivier Mattelaer (olivier-mattelaer)]

Dear Kilian,

Here is a simple fix:

=== modified file 'madgraph/interface/extended_cmd.py'
--- madgraph/interface/extended_cmd.py 2016-01-23 00:40:59 +0000
+++ madgraph/interface/extended_cmd.py 2016-02-23 16:48:44 +0000
@@ -579,7 +579,7 @@

         # Check if the line is complete
         if line.endswith('\\'):
- self.save_line = line[:-1]
+ self.save_line = line[:-1] + ' '
             return '' # do nothing

         # Remove comment

Thanks a lot for your help,

Olivier

> On Feb 23, 2016, at 15:49, Kilian Rosbach <email address hidden> wrote:
>
> Public bug reported:
>
> I've come across an unusual error message while running a simple
> MadGraph+MadSpin event generation (I believe the details are
> unimportant, but I can provide them if requested): MadSpin fails,
> complaining about an unknown particle "dl~dr~" (notice the missing space
> between what should actually be two SUSY particles).
>
> The exact message is: "InvalidCmd : No particle dl~dr~ in model"
>
> Studying the input files, I found the following continued (with \ at the end) line:
> "define susystrong = go ul ur dl dr cl cr sl sr t1 t2 b1 b2 ul~ ur~ dl~\
> dr~ cl~ cr~ sl~ sr~ t1~ t2~ b1~ b2~"
> in the file PROC_mssm_0/Cards/proc_card_mg5.dat
>
> It appears that the space following the backslash-newline combination is
> lost while reading this file. The problem disappears if I manually
> replace the file by one that doesn't use continued lines instead, but
> this is very difficult to do when running on the grid (the file is
> automatically generated by a script outside of my control).
>
> Is this easy to fix on your end?
>
> Thanks and best regards!
>
> ** Affects: mg5amcnlo
> Importance: Undecided
> Status: New
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1548882
>
> Title:
> problem with continuation lines in input files (probably in MadSpin
> code)
>
> Status in MadGraph5_aMC@NLO:
> New
>
> Bug description:
> I've come across an unusual error message while running a simple
> MadGraph+MadSpin event generation (I believe the details are
> unimportant, but I can provide them if requested): MadSpin fails,
> complaining about an unknown particle "dl~dr~" (notice the missing
> space between what should actually be two SUSY particles).
>
> The exact message is: "InvalidCmd : No particle dl~dr~ in model"
>
> Studying the input files, I found the following continued (with \ at the end) line:
> "define susystrong = go ul ur dl dr cl cr sl sr t1 t2 b1 b2 ul~ ur~ dl~\
> dr~ cl~ cr~ sl~ sr~ t1~ t2~ b1~ b2~"
> in the file PROC_mssm_0/Cards/proc_card_mg5.dat
>
> It appears that the space following the backslash-newline combination
> is lost while reading this file. The problem disappears if I manually
> replace the file by one that doesn't use continued lines instead, but
> this is very difficult to do when running on the grid (the file is
> automatically generated by a script outside of my control).
>
> Is this easy to fix on your end?
>
> Thanks and best regards!
>
> To manage n...

[Kilian Rosbach (kilian-rosbach)]

Hi Olivier,

thanks a lot for the super-fast reply!

I've looked at extended_cmd.py, and while I agree your fix will work for the particular case I mentioned, I think there might be an easy way to have a more general solution that will also work if the second line does not start with a space:

In lines 523-525, the first part of the continued line is added to the current line:
        if self.save_line:
            line = self.save_line + line
            self.save_line = ''
But just before that, in line 520, any spaces at the beginning of the line are removed:
        line = line.lstrip()

I would suggest to simply reverse the order of these two blocks of code, then it should work in all cases I believe.

Thanks again and cheers,
Kilian

[Olivier Mattelaer (olivier-mattelaer)]

Hi,

You are right this is a nicer fix.

Thanks a lot,

Olivier

> On Feb 23, 2016, at 17:54, Kilian Rosbach <email address hidden> wrote:
>
> Hi Olivier,
>
> thanks a lot for the super-fast reply!
>
> I've looked at extended_cmd.py, and while I agree your fix will work for
> the particular case I mentioned, I think there might be an easy way to
> have a more general solution that will also work if the second line does
> not start with a space:
>
> In lines 523-525, the first part of the continued line is added to the current line:
> if self.save_line:
> line = self.save_line + line
> self.save_line = ''
> But just before that, in line 520, any spaces at the beginning of the line are removed:
> line = line.lstrip()
>
> I would suggest to simply reverse the order of these two blocks of code,
> then it should work in all cases I believe.
>
> Thanks again and cheers,
> Kilian
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1548882
>
> Title:
> problem with continuation lines in input files (probably in MadSpin
> code)
>
> Status in MadGraph5_aMC@NLO:
> Fix Committed
>
> Bug description:
> I've come across an unusual error message while running a simple
> MadGraph+MadSpin event generation (I believe the details are
> unimportant, but I can provide them if requested): MadSpin fails,
> complaining about an unknown particle "dl~dr~" (notice the missing
> space between what should actually be two SUSY particles).
>
> The exact message is: "InvalidCmd : No particle dl~dr~ in model"
>
> Studying the input files, I found the following continued (with \ at the end) line:
> "define susystrong = go ul ur dl dr cl cr sl sr t1 t2 b1 b2 ul~ ur~ dl~\
> dr~ cl~ cr~ sl~ sr~ t1~ t2~ b1~ b2~"
> in the file PROC_mssm_0/Cards/proc_card_mg5.dat
>
> It appears that the space following the backslash-newline combination
> is lost while reading this file. The problem disappears if I manually
> replace the file by one that doesn't use continued lines instead, but
> this is very difficult to do when running on the grid (the file is
> automatically generated by a script outside of my control).
>
> Is this easy to fix on your end?
>
> Thanks and best regards!
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/mg5amcnlo/+bug/1548882/+subscriptions

[Kilian Rosbach (kilian-rosbach)]

Hello again,

it turns out that one additional change is needed to fully get things to work in the case of continued lines in proc_card_mg5.dat read by MadSpin. It took me a while to track this down, and maybe there is a more elegant:
In madgraph/various/banner.py, in the function read(self, init), the line = line.strip() statement needs to be removed, then it works for me.
In order to keep the change minimal, one could modify "self.append(store_line + line)" to "self.append((store_line + line).strip())" a few lines further down.

Cheers,
Kilian

[Olivier Mattelaer (olivier-mattelaer)]

Hi,

Thanks so much for the time that you spent to track down this. It is really appreciated.

Cheers,

Olivier

I will do the following:
=== modified file 'madgraph/various/banner.py'
--- madgraph/various/banner.py 2016-01-30 16:53:08 +0000
+++ madgraph/various/banner.py 2016-03-02 12:05:34 +0000
@@ -732,14 +732,17 @@

         store_line = ''
         for line in init:
- line = line.strip()
+ line = line.rstrip()
             if line.endswith('\\'):
                 store_line += line[:-1]
             else:
- self.append(store_line + line)
+ tmp = store_line + line
+ self.append(tmp.store_line)
                 store_line = ""
         if store_line:
             raise Exception, "WRONG CARD FORMAT"
+
+
     def move_to_last(self, cmd):
         """move an element to the last history."""
         for line in self[:]:

> On Mar 1, 2016, at 21:43, Kilian Rosbach <email address hidden> wrote:
>
> Hello again,
>
> it turns out that one additional change is needed to fully get things to work in the case of continued lines in proc_card_mg5.dat read by MadSpin. It took me a while to track this down, and maybe there is a more elegant:
> In madgraph/various/banner.py, in the function read(self, init), the line = line.strip() statement needs to be removed, then it works for me.
> In order to keep the change minimal, one could modify "self.append(store_line + line)" to "self.append((store_line + line).strip())" a few lines further down.
>
> Cheers,
> Kilian
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1548882
>
> Title:
> problem with continuation lines in input files (probably in MadSpin
> code)
>
> Status in MadGraph5_aMC@NLO:
> Fix Committed
>
> Bug description:
> I've come across an unusual error message while running a simple
> MadGraph+MadSpin event generation (I believe the details are
> unimportant, but I can provide them if requested): MadSpin fails,
> complaining about an unknown particle "dl~dr~" (notice the missing
> space between what should actually be two SUSY particles).
>
> The exact message is: "InvalidCmd : No particle dl~dr~ in model"
>
> Studying the input files, I found the following continued (with \ at the end) line:
> "define susystrong = go ul ur dl dr cl cr sl sr t1 t2 b1 b2 ul~ ur~ dl~\
> dr~ cl~ cr~ sl~ sr~ t1~ t2~ b1~ b2~"
> in the file PROC_mssm_0/Cards/proc_card_mg5.dat
>
> It appears that the space following the backslash-newline combination
> is lost while reading this file. The problem disappears if I manually
> replace the file by one that doesn't use continued lines instead, but
> this is very difficult to do when running on the grid (the file is
> automatically generated by a script outside of my control).
>
> Is this easy to fix on your end?
>
> Thanks and best regards!
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/mg5amcnlo/+bug/1548882/+subscriptions

[Kilian Rosbach (kilian-rosbach)]

Hi Olivier,

I'm happy to help!

I agree it's a good to add the rstrip() command in case there are any
newline characters present.

If I'm not mistaken you probably wanted to write
self.append(tmp.strip())
instead of
self.append(tmp.store_line)

Do you perhaps know already when this will appear in an official
MadGraph release?

Thanks and cheers,
Kilian

On 02.03.2016 13:08, Olivier Mattelaer wrote:
> Hi,
>
> Thanks so much for the time that you spent to track down this. It is
> really appreciated.
>
> Cheers,
>
> Olivier
>
> I will do the following:
> === modified file 'madgraph/various/banner.py'
> --- madgraph/various/banner.py 2016-01-30 16:53:08 +0000
> +++ madgraph/various/banner.py 2016-03-02 12:05:34 +0000
> @@ -732,14 +732,17 @@
>
> store_line = ''
> for line in init:
> - line = line.strip()
> + line = line.rstrip()
> if line.endswith('\\'):
> store_line += line[:-1]
> else:
> - self.append(store_line + line)
> + tmp = store_line + line
> + self.append(tmp.store_line)
> store_line = ""
> if store_line:
> raise Exception, "WRONG CARD FORMAT"
> +
> +
> def move_to_last(self, cmd):
> """move an element to the last history."""
> for line in self[:]:
>
>
>> On Mar 1, 2016, at 21:43, Kilian Rosbach <email address hidden> wrote:
>>
>> Hello again,
>>
>> it turns out that one additional change is needed to fully get things to work in the case of continued lines in proc_card_mg5.dat read by MadSpin. It took me a while to track this down, and maybe there is a more elegant:
>> In madgraph/various/banner.py, in the function read(self, init), the line = line.strip() statement needs to be removed, then it works for me.
>> In order to keep the change minimal, one could modify "self.append(store_line + line)" to "self.append((store_line + line).strip())" a few lines further down.
>>
>> Cheers,
>> Kilian
>>
>> --
>> You received this bug notification because you are subscribed to
>> MadGraph5_aMC@NLO.
>> https://bugs.launchpad.net/bugs/1548882
>>
>> Title:
>> problem with continuation lines in input files (probably in MadSpin
>> code)
>>
>> Status in MadGraph5_aMC@NLO:
>> Fix Committed
>>
>> Bug description:
>> I've come across an unusual error message while running a simple
>> MadGraph+MadSpin event generation (I believe the details are
>> unimportant, but I can provide them if requested): MadSpin fails,
>> complaining about an unknown particle "dl~dr~" (notice the missing
>> space between what should actually be two SUSY particles).
>>
>> The exact message is: "InvalidCmd : No particle dl~dr~ in model"
>>
>> Studying the input files, I found the following continued (with \ at the end) line:
>> "define susystrong = go ul ur dl dr cl cr sl sr t1 t2 b1 b2 ul~ ur~ dl~\
>> dr~ cl~ cr~ sl~ sr~ t1~ t2~ b1~ b2~"
>> in the file PROC_mssm_0/Cards/proc_card_mg5.dat
>>
>> It appears that the space following the backslash-newline combination
>> is lost while reading this file. The problem disappears if I manually
>> replace the fil...

[Olivier Mattelaer (olivier-mattelaer)]

Hi Kilian,

> If I'm not mistaken you probably wanted to write
> self.append(tmp.strip())
> instead of
> self.append(tmp.store_line)

Oops, I realized my mistake and change it in the code, but the diff was already created and I forget to edit it.
Sorry.

> Do you perhaps know already when this will appear in an official
> MadGraph release?

We are preparing the 2.4.0 version for the moment. All the feature that we want to put it are in the internal review process.
So I hope it will out in before the end of the month at latest.

Cheers,

Olivier

> On Mar 2, 2016, at 15:46, Kilian Rosbach <email address hidden> wrote:
>
> Hi Olivier,
>
> I'm happy to help!
>
> I agree it's a good to add the rstrip() command in case there are any
> newline characters present.
>
> If I'm not mistaken you probably wanted to write
> self.append(tmp.strip())
> instead of
> self.append(tmp.store_line)
>
> Do you perhaps know already when this will appear in an official
> MadGraph release?
>
> Thanks and cheers,
> Kilian
>
> On 02.03.2016 13:08, Olivier Mattelaer wrote:
>> Hi,
>>
>> Thanks so much for the time that you spent to track down this. It is
>> really appreciated.
>>
>> Cheers,
>>
>> Olivier
>>
>> I will do the following:
>> === modified file 'madgraph/various/banner.py'
>> --- madgraph/various/banner.py 2016-01-30 16:53:08 +0000
>> +++ madgraph/various/banner.py 2016-03-02 12:05:34 +0000
>> @@ -732,14 +732,17 @@
>>
>> store_line = ''
>> for line in init:
>> - line = line.strip()
>> + line = line.rstrip()
>> if line.endswith('\\'):
>> store_line += line[:-1]
>> else:
>> - self.append(store_line + line)
>> + tmp = store_line + line
>> + self.append(tmp.store_line)
>> store_line = ""
>> if store_line:
>> raise Exception, "WRONG CARD FORMAT"
>> +
>> +
>> def move_to_last(self, cmd):
>> """move an element to the last history."""
>> for line in self[:]:
>>
>>
>>> On Mar 1, 2016, at 21:43, Kilian Rosbach <email address hidden> wrote:
>>>
>>> Hello again,
>>>
>>> it turns out that one additional change is needed to fully get things to work in the case of continued lines in proc_card_mg5.dat read by MadSpin. It took me a while to track this down, and maybe there is a more elegant:
>>> In madgraph/various/banner.py, in the function read(self, init), the line = line.strip() statement needs to be removed, then it works for me.
>>> In order to keep the change minimal, one could modify "self.append(store_line + line)" to "self.append((store_line + line).strip())" a few lines further down.
>>>
>>> Cheers,
>>> Kilian
>>>
>>> --
>>> You received this bug notification because you are subscribed to
>>> MadGraph5_aMC@NLO.
>>> https://bugs.launchpad.net/bugs/1548882
>>>
>>> Title:
>>> problem with continuation lines in input files (probably in MadSpin
>>> code)
>>>
>>> Status in MadGraph5_aMC@NLO:
>>> Fix Committed
>>>
>>> Bug description:
>>> I've come across an unusual error message while running a simple
>>> MadGraph+MadSpin event generation...