MapNode fails to aggregate dynamic outputs of DynamicTraited interfaces

[oesteban]

It is easy to reproduce with a JSONFileGrabber inside a MapNode, for instance.

[stebo85]

Is there any progress on this bug? I think I am getting it when I try to do this:
params_n = MapNode(JSONFileGrabber(),
name='parameter_node', iterfield=['in_file'])

Traceback (most recent call last):
File "/QRISdata/Q0538/17042_detection_of_concussion/tools/miniconda3/lib/python3.6/site-packages/nipype/pipeline/engine/nodes.py", line 535, in _get_inputs
output_value = results.outputs.trait_get()[output_name]
AttributeError: 'Bunch' object has no attribute 'trait_get'

It works when I use a normal Node, but not a MapNode - any ideas how I can get around this problem?

[stebo85]

This is my current workaround in case someone else has a problem with using the JSONFileGrabber in a MapNode:

def read_json(in_file):
    import os
    te = 0.001
    b0 = 7
    if os.path.exists(in_file):
        import json
        with open(in_file, 'rt') as fp:
            data = json.load(fp)
            te = data['EchoTime']
            b0 = data['MagneticFieldStrength']

    return te, b0


params_n = MapNode(interface=Function(input_names=['in_file'],
                                      output_names=['EchoTime', 'MagneticFieldStrength'],
                                      function=read_json),
                   name='read_json', iterfield=['in_file'])

wf.connect([(selectfiles, params_n, [('params', 'in_file')])])