cross py2/py3 python object serializer, currently using and extending jsonpickle
to dump/load python objects with a primary focus on numpy/pandas types.
The main purpose of numpyson is to allow dumping under a python3 interpreter
and loading under python2 and vice versa. This is a use case that is not covered
by current serializers AFAIK.
quick simple API example:
s = numpyson.dumps(numpy.array([1,2,3]))
data = numpyjson.loads(s)
assert data.to_list() == [1,2,3]
- numpy arrays and some other numpy types
- nested python data structures with the above types
- fix bug: respect ordering when serializing/deserializing ndarrays
- depend on pandas>=0.13.1 because we are using Float64Index which
does not exist in pandas-0.12
- support numpy.float64 and numpy.int64
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.