Setting the Command Set Property Values in N_CREATE and other responses

[sscotti]

Not sure how to do this. Setting the tags show below for the N-CREATE-RSP in the modified attr_list is not the proper way to do that, but I don't see how to explicitly set those in an N_CREATE handler ?

 'N-CREATE-RSP': (
         'CommandGroupLength', 'AffectedSOPClassUID', 'CommandField',
         'MessageIDBeingRespondedTo', 'CommandDataSetType', 'Status',
         'AffectedSOPInstanceUID',
         'ErrorID', 'ErrorComment'
),

I've been testing a MPPS server as outlined in the examples and put some debugging info in there, e.g.:

def handle_create(event):
    print(event.event.description) # InterventionEvent(name='EVT_N_CREATE', description='N_CREATE request received')
    print(event.timestamp) #2021-07-04 00:29:59.168226
    print("Message ID:  " + str(event.message_id)) # Unique number per client, int
    print('Context ID:  ' + str(event.context.context_id)) # Unique number per client, int
    print('Abstract Syntax:  ' + str(event.context.abstract_syntax)) # Unique number per client, int
    print('Transfer Syntax:  ' + str(event.context.transfer_syntax)) # Unique number per client, int
    print('AE:  ' + str(event.assoc.ae.ae_title)) # Unique number per client, int

python_mpps_1         | N-CREATE request received
python_mpps_1         | 2021-07-04 17:45:08.034944
python_mpps_1         | Message ID:  9
python_mpps_1         | Context ID:  1
python_mpps_1         | Abstract Syntax:  1.2.840.10008.3.1.2.3.3
python_mpps_1         | Transfer Syntax:  1.2.840.10008.1.2
python_mpps_1         | AE:  b'MPPS_SERVER 

I think the Command Set has to be set is some way other than just adding them to the dataset, but as a Command Set Object ? I do not know how to do that. The reason I say that is because I've been testing with DVTk, and added the Command Set tags to the returned dataset, and DVTk does not pick them up, probably because it expects to see them in another location.

I just started playing around with this package, so excuse the ignorance. I nice way to sort of learn though is to examine objects in a fashion sort of like above.

Kind of wondering if the DIMSEServiceProvider or Association has to be used to add the needed items to the Command Set for the N_CREATE response. I have no idea what CommandGroupLength, CommandField and CommandDataSetType are supposed to be. I think I can figure out the other ones mostly.

Thanks.

[scaramallion]

You shouldn't need to, they're set automatically from the corresponding message primitive. Is there some reason you need to customise them?

For N-CREATE you only need to return the operation's Status and the returned Attribute List dataset (if any). All other elements in the DIMSE response are set automatically.

[sscotti]

OK, I was hoping so. It might be have something to do with the validation setup I have with my DVTk validation script. I tried it in a live environment with an MRI scanner and it could C_ECHO just fine, and the rest of the N_CREATE script worked, but in DVTk I am getting a warning regarding the response. The scanner was throwing an error with the N_CREATE, but I think I have not tested it yet again after I made some changes.

RESULT: PASSED

Thanks.
Number of Validation Errors: 0 - Number of Validation Warnings: 0
Number of User Validation Errors: 0 - Number of User Validation Warnings: 0
Number of General Errors: 0 - Number of General Warnings: 1

Warning: Unexpected DICOM Dataset received - DICOMScript does not match Product behaviour.

I'll have to recheck everything and try it again in the live environment.

I pulled the basic framework from this example: https://pydicom.github.io/pynetdicom/stable/examples/mpps.html#mpps-scp

[scaramallion]

The DIMSE-N examples are based off my reading of the DICOM Standard, so it's certainly possible that the examples themselves may not be entirely correct. I'd try and get the DICOM Conformance statement for the MR as well, in case they're expecting something that's otherwise optional.

[sscotti]

I'll have the opportunity to test things again in a live environment soon with better logging of the comms between an MRI scanner and the MPPS server, so I'll be able to debug things better if there is still some sort of communication error. I'll have to see also if there is a way with DVTk to capture or display the returned Command Set, which would be helpful. That is a nice set of tools (and free). The only drawback is that it runs on Windows. https://www.dvtk.org/ They have an extensive script testing library, which is really excellent for development.