Effectively disable LOQReductionGUI systemtest

[Anders Markvardsen]

LOQReductionGUI systemtest is failing with

​http://builds.mantidproject.org/job/develop_systemtests_rhel6/605/

Of interest is that there has been not commit made to ISIS SANS for commit made which causes this to fail.

Get build server back and make some understanding of why this test is failing my some change otherwise in Mantid not related to ISIS SANS.

[Anders Markvardsen]

increase tolerance for logreductiongui. re #10287

Changeset: bda4a07603669e73373e7d90f6ff7da0a373f5c3

[Anders Markvardsen]

To test:

1) test that LOQReductionGUI is not failing on ​http://builds.mantidproject.org/job/develop_systemtests_rhel6/

2) optionally run the LOQReductionGUI system and check that by eye the output and reference output basically overlap

Note although by eye the change looks ok, I have created #10288 for a rainy day and with instrument scientist to give the ok

[Anders Markvardsen]

Note for reference. This build stopped working from ​http://builds.mantidproject.org/job/develop_systemtests_rhel6/602/ with the changes in dependencies: 1.rhel6-build #1958 → #1963 (detail)

[Anders Markvardsen]

will give git by-sect a go a recommended by Martyn

[Anders Markvardsen]

Observations from locating in more detail why this systemtest is failing:

Checking out develop main code branch and systemtest master branch. From develop main code branch the following commit works (​http://builds.mantidproject.org/job/develop_incremental/1958/):

git checkout 75778cdef4fb43e4a825348e22e2ba0f81bc81df

and it is the next commit (​http://builds.mantidproject.org/job/develop_incremental/1959/) which brakes the LOQ systemtest for this ticket. Check this branch out with:

git checkout d0c9eedb57e75edc7bf2ad8f7a724d7a637899ab

This is caused by changes to TabulatedFunction done in #10279. This tickets add new fitting parameter Centre to this function.

This function is used by SANS reduction in script/SANS/ISISCommandInterface.py in lines

{ (573 ) Function='name=TabulatedFunction, Workspace="'+str(frontData)+'"' (578 ) Function='name=TabulatedFunction, Workspace="'+str(frontData)+'"' (583 ) function_input = 'name=TabulatedFunction, Workspace="'+str(frontData)+'"' +";name=FlatBackground" (588 ) Function='name=TabulatedFunction, Workspace="'+str(frontData)+'"' (593 ) Function='name=TabulatedFunction, Workspace="'+str(frontData)+'"' (599 ) Function='name=TabulatedFunction, Workspace="'+str(frontData)+'"' (603 ) Fit(InputWorkspace=rearData, Function='name=TabulatedFunction, Workspace="'+str(frontData)+'"' }

I tried to change the code in ISISCommandInterface.py to ties Centre=0.0, but when this code is used in the content of running the systemtest LOQReductionGUI, at least from my attempt, it still fails because the TabulatedFunction is failing.

[Anders Markvardsen]

(In #10279) Hi Jose,

I have put this ticket as blocked by #10287.

The reason is that your changes have caused a ISIS SANS LOQ systemtest to fail. Please read comment 7 of this ticket, it has been found that your changes to TabulatedFunction is unfortunately now causing TabulatedFunction to fail, see comment 7 in #10287.

Could you either update the code referred to in comment 7 of #10287 or advice me on how I can get TabulatedFunction to work again in the code referred to in this comment 7

[Jose Borreguero]

If you want to reproduce the previous behavior, you can introduce a tie that will fix the new fitting parameter 'Shift' of TabulatedFunction to the default value:

Ties='f0.Shift=0.0'

You can combine ties if you also want to fix the 'Scaling' parameter:

Ties='f0.Scaling={0},f0.Shift=0.0'.format(rAnds.scale)

I am not sure but you may need to insert extra parenthesis:

Ties='(f0.Scaling={0},f0.Shift=0.0)'.format(rAnds.scale)

Finally, I don't know about the physics of this problem, but you may also want to take advantage of the Shift parameter now that it is available :)

[Anders Markvardsen]

Hi Jose,

I notice in d0c9eedb57e75edc7bf2ad8f7a724d7a637899ab that you have added a new parameter Centre not Shift.

Did you mean to write f0.Centre=0.0 instead of f0.Shift=0.0 ?

[Jose Borreguero]

I just commited ​/ a1bccc219b35f726b8185f8051b5a5d0b39a0a89 where I changed the name to Shift. It made more sense.

[Jose Borreguero]

I have finished with ticket #10279. Can you please remove the block so that I can close it ?

[Anders Markvardsen]

Hi Jose,

Please note that I checked as mentioned in comment 7 with Centre (as it was called at that point) fixed and I still got a different fit, which I was surprissed by.

I will try to do this again with the most recent version of your code.

[Jose Borreguero]

Hi Anders,

Were you able to fix the fitting? If so, can you please remove the block from #10279 ?

[Anders Markvardsen]

Hi Jose,

I see your point, track appears to not have updated blocking and blocked by as I expected, i.e. track had 10287 blocking 10279 and 10279 blocking 10287 which is a stalemate! Now this ticket is blocked by 10287.

[Anders Markvardsen]

Hi Jose,

Note trac appears to have gone ga ga.

I have now tried to entire remove any block with 10279 and 10287.

See if you can put 10287. If not try to see if you can remove the block.

[Jose Borreguero]

You could give me temporary ownership of this ticket, and then I'll try to remove this ticket from blocking #10279. It may work if I own both tickets.

[Jose Borreguero]

Hi Anders,

I took ownership of the ticket in order to see if I could remove the block, but couldn't, so I returned ownership to you.

Anyway, I understand the system test is working again, so you could flag the ticket as fixed, and maybe after the ticket has been reviewed and closed the block will be lifted. What do you think?

[Anders Markvardsen]

Hi Jose,

OK.

To tester: please push changes systemtest to master, which effectively temperaroty disable this test. This will be reverted in #10288 after #10279 have been pushed to master.

[Jose Borreguero]

If I understood correctly, I only have to "git test pass" because the implemented changes will disable the offending test, right?

-Jose

[Stuart Campbell]

This ticket has been transferred to github issue ​11129