allow the insertion of violated scans

[gluneau]

This command will insert a violated scan since providing the acquisition protocol bypasses the protocol checks. If you also need to bypass the extra_file_checks use bypass_extra_file_checks
Please note the -create_minc_pics option as well since it is usually created by the tarchiveloader script.

minc_insertion.pl -acquisition_protocol t2w -bypass_extra_file_checks -create_minc_pics -profile prod -globLocation -force  -tarchivePath /Users/xxxproject/dataTransfer/library/2009/DCM_2009-09-25_project_20110214_185904581.tar -mincPath /data/project/data/trashbin/TarLoad-3-34-pVzGC5/xxx0067_703739_v12_20090925_222403_18e1_mri.mnc

[MounaSafiHarab]

@gluneau,

you want to create pics if the user spoecifies this option; this will do the opposite
(I just confirmed when running the command line: pics generated withOUT specifying -create_minc_pics)

[MounaSafiHarab]

@gluneau,

when running the command we should not specify -create_minc_pics 1; it should simply be -create_minc_pics (without the 1)

[MounaSafiHarab]

@gluneau,

$acquisitionProtocol is no longer accessible for the next check $utility->registerScanIntoDB() (line 416)

[driusan]

The help should describe what the flag does, not when to use it. ("without" is also spelled wrong.)

[MounaSafiHarab]

@gluneau;

I think:

1. you forgot to add the MRI.pm file to your last commit (which has the main create_minc_pics());
   and
2. tarchiveLoader should be updated as well by removing the call to mass_pic.pl

[gluneau]

@MounaSafiHarab

1. The MRI.pm call was already present in the file see other make_ function calls. (line 16)

2. The tarchieveLoader modification is in this request: Move minc pics creation to MRI.pm #149

[MounaSafiHarab]

@gluneau

This has to be tested together with the changes in: #149
otherwise "Undefined subroutine &NeuroDB::MRI::make_minc_pics called at uploadNeuroDB/minc_insertion.pl line 473."

Otherwise, it looks good.

(a minor comment: I would modify the comment as to how to run it by removing the leading 'perl' before the rest of the command, as this is how tarchiveLoader script runs minc_insertion.pl)

[gluneau]

@MounaSafiHarab running instruction updated to not have the perl program in front.

[driusan]

Acquisition_protocol as a parameter should only specify the acqusition_protocol, not also bypass the protocol enforcement.

There should be two flags:

1. -acquisition_protocol (to specify the acquisition protocol if the mri_protocol table can't determine it)
2. -force (to skip the checks and insert it regardless of the study protocol)

[MounaSafiHarab]

@driusan @gluneau

the -force flag is already used when checking if the isTarchiveValidated flag is set in the mri_upload table, so I suggested to Greg to use another flag that is indicative of forcing specifically no extra_file_check. I assume you are Okay with this @driusan ?

[driusan]

Do the flags get passed between the scripts? If not, I'm not sure it would be a problem and -force is a pretty standard name across different software packages. But if there's a conflict or you think it would be confusing, I have no problem using a different name (-forceinsert, maybe?)

[MounaSafiHarab]

Currently our imaging_upload_file.pl which runs all the insertion scripts calls tarchiveLoader withOUT the -force option. As we have it now, the -force is used to bypass tarchive validation checks in tarchiveLoader and bypass the isTarchiveValidated flag check in mri_upload table in the minc_insertion script.

I think we have two options:

1. If it is to be used as is (i.e. -force), then a) the help section which describes the use of this flag needs to be updated to say that we are bypassing tarchiveValidation and extra_file_checks(), and b) projects that might have customized this flag for their study need to be made aware of this
2. Use a flag name that is more specific to what we are doing: namely bypassing extra_file_checks() (as suggested by @gluneau) or -forceinsert (as syuggested by @driusan)

I personally prefer option (2)! it is more flags, but it is clearer I think!

[MounaSafiHarab]

@gluneau

tested it on a file that would violate the mri_protocol_checks entry, with and without the -bypass_extra_file_checks flag, and it seems to be NOT inserting without the flag and inserting with the flag on. So I think this is all good now.