Testing I/O of JSONs

for eventually creating config files to drive fMRI analysis



In [2]:

    
import json



In [3]:

    
con = 'cdmainrel'
tt='CD'
debug = True
write = False
roi = 'grayMatter'
dstype = 'raw'
r = 4  # searchlight radius
sg_params = [119, 2]



In [4]:

    
analysisConfig = {'con': 'cdmainrel', 'tt': 'CD', 'write': 'False', 'roi': 'grayMatter', 'dstype': 'raw', 'r': 4, 'sg_params': [119, 2]}



In [11]:

    
json.dumps([analysisConfig, analysisConfig, analysisConfig])









    Out[11]:





'[{"roi": "grayMatter", "sg_params": [119, 2], "dstype": "raw", "tt": "CD", "write": "False", "r": 4, "con": "cdmainrel"}, {"roi": "grayMatter", "sg_params": [119, 2], "dstype": "raw", "tt": "CD", "write": "False", "r": 4, "con": "cdmainrel"}, {"roi": "grayMatter", "sg_params": [119, 2], "dstype": "raw", "tt": "CD", "write": "False", "r": 4, "con": "cdmainrel"}]'



In [12]:

    
newAnalysisConfig = json.loads(json.dumps([analysisConfig, analysisConfig, analysisConfig]))



In [13]:

    
newAnalysisConfig[0]









    Out[13]:





{u'con': u'cdmainrel',
 u'dstype': u'raw',
 u'r': 4,
 u'roi': u'grayMatter',
 u'sg_params': [119, 2],
 u'tt': u'CD',
 u'write': u'False'}

Seems like we can just save all the arguments as a dict of dicts... easy!

Necessary fields:

Analysis parameters
File paths
Write paths
Complicated stuff (like cross validation, etc...) solve with pickle? how to set these up? with Jupyter notebook?



In [8]:

    
type(analysisConfig) == dict









    Out[8]:





True



In [ ]: