PiWind example fails on Python 3.6 because of dictionary comparisons about oasislmf HOT 2 CLOSED

@ElliotJH Yes, this came up during testing. In fact, in the original version of the code six wasn't being used, as any dict has itervalues. You can do the following

tiv_elements = tuple(sorted(
                [v for v in cep.itervalues() if v.get('FieldName') and v['FieldName'] == 'TIV'],
                key=lambda v: v['FMTermGroupID']
            ))

The sort is required because the point is to extract an list of TIV elements/fields (these are dict values in cep describing TIV elements which correspond to TIV fields, and these TIV fields describe column names in a CSV exposures file). The point is that the output of this line should be a list of TIV element dicts ordered element names, so for PiWInd the element for WSCV1VAL should be listed before the one for WSCV4VAL, etc.

from oasislmf.

@ElliotJH It appears I was wrong - Python 3.6 dicts don't have itervalues, instead have values which appear to be generators. I mainly develop in Python 2.7, and am less familiar with 3+. So the issue you pointed out is still valid. In fact the solution is as follows

tiv_elements = tuple(
    sorted(
        [v for v in canonical_exposures_profile.values() if v.get('FieldName') == 'TIV'],
        key=lambda v: v['FMTermGroupID']
    )
)

So filter in Python 3 produces a generator not a list as in Python 2, so your solution wouldn't work.Tthe common solution is to replace the filter with a plain list, and loop over <dict>.values() not <dict>.itervalues().

This will be fixed in the next release.

from oasislmf.