VTRTool

Python utilities for Fullwave3D VTR model files.

The package contains a vtrtool executable, as well as a vtrtool.py python package

Installation

This tool can be obtained from PyPI. Simply do:

pip install vtrtool

vtrtool executable quick usage

vtrtool -h to show usage information

vtrtool info my_model_file.vtr shows metadata info for my_model_file.vtr

vtrtool info nx1 my_model_file.vtr prints out just the size of the 1st (x) dimension for my_model_file.vtr

vtrtool to_binary --property=2 my_model_file.vtr my_model_file.bin to dump the 2nd property from my_model_file.vtr to the float32 raw binary file my_model_file.bin

vtrtool to_stdout my_model_file.vtr to pipe the 1st property (default) from my_model_file.vtr to stdout in raw float32 binary format

vtrtool.py Python module quick usage

The vtrtool.VTRModel class represents a vtr model that lives at a given filename. The file will be parsed on object initialization:

import vtrtool

# read the vtr file `my_model_file.vtr` and construct a new VTRModel object with it
vtr_model = vtrtool.VTRModel("my_model_file.vtr")

# print model properties to stdout
vtrtool.print_vtr_metadata(vtr_model)

# model property 1 will be available as numpy ndarray in
vtr_model.arrays[0]

# model property 2 (if it exists) will be available as numpy ndarray in
vtr_model.arrays[1]

# etc...

Set the optional keyword use_memmap=True if you want the VTRModel object to use a NumPy memmap instead of in-memory representation for holding the model array.

The following instance properties exist for objects of the VTRModel class:

• file: (str) the vtr filename referenced by the VTRModel object
• num_dimensions: (int) number of dimensions in the model
• num_attributes: (int) number of properties in the model
• shape: (tuple of ints) shape (c-based) of numpy ndarray holding the models in the order (nx1, nx2, nx3)
• shape_fortran: (tuple of ints) shape (fortran-based) of the models in the order (nx3, nx2, nx1)
• nx1: (int) number of cells in the 1st (usually x) dimension
• nx2: (int) number of cells in the 2nd (usually y) dimension
• nx3: (int) number of cells in the 3rd (fastest, usually z) dimension
• arrays: (list of ndarrays) holds the model property ndarrays, the length of the list correspond to the number of model properties (i.e., property 1 is in model[0]), and the shape of the ndarray will correspond to shape and with the number of dimensions in num_dimensions

Some convenience functions:

• vtrtool.vtrfile_to_ndarray(filename, prop_index=0) quickly takes a vtr file and returns the property specified by prop_index as an ndarray
• vtrtool.print_vtr_metadata(vtrmodel_object) takes a VTRModel object and prints dimension information to stdout

Maintained by

Tim Lin tlin@s-cube.com, 2017