todo

Mar 18 2016

HDF5 1.10 released, 
JNI can be compiled, but how can I use MPI to parse the non-mpi hdf5 calls. 

https://www.hdfgroup.org/ftp/HDF5/releases/hdf5-1.10/hdf5-1.10.0-pre1/src/hdf5-1.10.0-pre1-RELEASE.txt

Jan 5 2016
1. install parallel h5py inside anaconda on mac
2. install parallel hdf5 on mac
3. on cori, test mpi inside spark, by initializing spark first then init mpi, or spawn child processes



Mar 2 2016
    #to do list:
	#check file size,
        # if small, sequential read
        # if large, parallel read
        # on default, sequential read


# def __init__(self,filename,datasetname, cooridnates, conditions):
#     self.fname=filename
#     self.dname=datasetname
#     self.cord=cooridnates
#     self.cond=conditions
#read by coordinates, e.g., start:end=0:100

#read data by given parameters: x[0]file name,x[1]dataset name


       #f.close()
#def rangesplit(x):

#read by conditions, val<10
#def readbycond(self):
#    f=h5py.File(self.fname,'r')
#    d=f[self.dname]
#    return list(d[:])