fixed fsl.Level1Design gamma convoluion

[TheChymera]

This also provides a framework that is easily extensible to other convolution methods which we do not yet provide.

Fixes issue #1498

[coveralls]

Coverage increased (+0.003%) to 72.076% when pulling f28665b on TheChymera:ld into 124a850 on nipy:master.

[satra]

thanks @TheChymera - could we add two more pieces?

1. an updated CHANGES file
2. a small test for the function that ensures that the hrf is set to the correct basis_key. the test could create a fake condition.

[TheChymera]

1. where do I add a line to the CHANGES file?

Presumably not under the following header - right? Do I create a new header?

Release 0.12.0-rc1 (April 20, 2016)
============

2. Can you provide an example of how such a test should be formatted and where it should be put? On a side note, this might be a bit difficult - as the Level1Design node requires a session_info argument, which has to contain a path to an existing 4D NIfTI file. nipype doesn't ship with any of those - right?

[satra]

@TheChymera - under Release 0.12.0-rc1 is fine, we will be updating that for the full release soon.

re: the test, i was thinking that the test could go into ../interfaces/fsl/tests/test_model

this test would specifically use the _create_ev_files function of the interface, for which all i think you need to do is sth along the lines of:

from ...base import undefined
runinfo = dict(cond=[{'name': 'test_condition', 'onsets': [0, 10], durations=[10, 10]}])
runidx = 0
contrasts = undefined
usetd = False
do_tempfilter = False

basis_key will change for each type of hrf

and then check that the return from the function contains the right basis_key.

this doesn't check the actual interface, which will be a separate repo of tests we will be developing.

[satra]

ev_none is still needed for the regressors

[coveralls]

Coverage increased (+0.02%) to 72.089% when pulling 79c71f1 on TheChymera:ld into 124a850 on nipy:master.

[TheChymera]

@satra I can't test my test script because no matter how I call it I get ValueErrors somewhere down the line due to the relative imports, e.g.:

chymera@quiethost ~/src/nipype/nipype/interfaces $ python -m fsl.tests.test_Level1Design.test_level1design
Traceback (most recent call last):
  File "/usr/lib64/python2.7/runpy.py", line 151, in _run_module_as_main
    mod_name, loader, code, fname = _get_module_details(mod_name)
  File "/usr/lib64/python2.7/runpy.py", line 101, in _get_module_details
    loader = get_loader(mod_name)
  File "/usr/lib64/python2.7/pkgutil.py", line 464, in get_loader
    return find_loader(fullname)
  File "/usr/lib64/python2.7/pkgutil.py", line 474, in find_loader
    for importer in iter_importers(fullname):
  File "/usr/lib64/python2.7/pkgutil.py", line 430, in iter_importers
    __import__(pkg)
  File "fsl/__init__.py", line 9, in <module>
    from .base import (FSLCommand, Info, check_fsl, no_fsl, no_fsl_course_data)
  File "fsl/base.py", line 28, in <module>
    from builtins import object
  File "/usr/lib64/python2.7/site-packages/builtins/__init__.py", line 8, in <module>
    from future.builtins import *
  File "/usr/lib64/python2.7/site-packages/future/builtins/__init__.py", line 13, in <module>
    from future.builtins.misc import (ascii, chr, hex, input, isinstance, next,
  File "/usr/lib64/python2.7/site-packages/future/builtins/misc.py", line 43, in <module>
    from io import open
  File "io.py", line 38, in <module>
    from .base import (TraitedSpec, traits, File, Directory,
ValueError: Attempted relative import in non-package

Hoe do you do it?

[satra]

you can use:

nosetests nipype.interfaces.fsl.tests.test_maths:test_meanimage

but this requires nipype to be installed (python setup.py install or python setup.py develop)

[satra]

the current updates still use ev_none here:

https://github.com/nipy/nipype/pull/1500/files#diff-eade534f1c2e76b62be0e071aae3324cR195

[TheChymera]

the current updates still use ev_none here:

Yes, isn't that supposed to happen for the regressors? I never used this feature, so I didn't even realize ev_none was still needed before you pointed it out. Now I reintroduced it.

[coveralls]

Coverage increased (+0.02%) to 72.088% when pulling 34c43e6 on TheChymera:ld into 124a850 on nipy:master.

[satra]

this doesn't really test it. the following would run a test. something would need to be replaced with the expected string

output = Level1Design._create_ev_files(l,os.getcwd(),runinfo,runidx,usetd,contrasts,do_tempfilter,"none")
yield assert_true, 'something' in output
output = Level1Design._create_ev_files(l,os.getcwd(),runinfo,runidx,usetd,contrasts,do_tempfilter,"dgamma")
yield assert_true, 'something' in output
output = Level1Design._create_ev_files(l,os.getcwd(),runinfo,runidx,usetd,contrasts,do_tempfilter,"gamma")
yield assert_true, 'something' in output

[coveralls]

Coverage increased (+0.1%) to 72.201% when pulling f75efb6 on TheChymera:ld into 124a850 on nipy:master.

[TheChymera]

@satra could you please pul?