HDF5 Tools

Different tools to work with datasets in HDF5 file format.

Created on Thu Apr 2 21:12:46 2015

@author: akusok

hpelm.modules.hdf5_tools.make_hdf5(data, h5file, dtype=<type 'numpy.float64'>, delimiter=' ', skiprows=0, comp_level=0)

Makes an HDF5 file from whatever given data.

Parameters:

• data –
  • input data in Numpy.ndarray or filename, or a shape tuple
• h5file –
  • name (and path) of the output HDF5 file
• delimiter –
  • data delimiter for text, csv files
• comp_level –
  • compression level of the HDF5 file

hpelm.modules.hdf5_tools.normalize_hdf5(h5file, mean=None, std=None, batch=None)

Calculates and applies normalization to data in HDF5 file.

Parameters:

• mean –
  • known vector of mean values
• std –
  • known vector of standard deviations
• batch –
  • number of rows to read at once, default is a native batch size