Summary
In some objects, tuples, namedtuples, OrderedDict and other data-container objects are used. Currently, MontyEncoder and MontyDecoder do not support them (it actually converts tuples to lists, OrderedDicts to dicts)
- When decoding back these objects, they are not anymore the same type of data object and this could be a problem in some cases.
- For OrderedDict, it might not be such of a problem as dicts are now guaranteed to be ordered in python 3.7, if I'm not wrong.
Is there any plan to implement such a feature ? Would it be useful for people ?
Example code for namedtuple
from collections import namedtuple
from monty.json import MontyEncoder
from monty.json import MontyDecoder
me = MontyEncoder()
md = MontyDecoder()
Point = namedtuple('Point', field_names=['x', 'y'])
a = Point(1, 2)
mydict = {'point': a}
mydecodeddict = md.decode(me.encode(mydict))
print(mydict == mydecodeddict) # Returns FalseSuggested solution (if known)
- Deal with these specific data objects in a similar way numpy is handled.
Summary
In some objects, tuples, namedtuples, OrderedDict and other data-container objects are used. Currently, MontyEncoder and MontyDecoder do not support them (it actually converts tuples to lists, OrderedDicts to dicts)
Is there any plan to implement such a feature ? Would it be useful for people ?
Example code for namedtuple
Suggested solution (if known)