- The major change here was replacing numarray with numpy.
For general instructions on converting scripts to use numpy see http://www.stsci.edu/resources/software_hardware/numarray/numarray2numpy.pdf.
The specific changes to esys.escript are:
- getValueOfDataPoint() which returned a numarray.array has been replaced by
getTupleForDataPoint() which returns a python tuple containing
the components of the data point. In the case of matricies or higher ranked data, the tuples will be nested. Use
numpy.array(data.getTupleForDataPoint()) if a numpy.ndarray object is required.
- getValueOfGlobalDataPoint has similarly been replaced by getTupleForGlobalDataPoint().
- integrate(data) now returns a numpy.ndarray instead of a numarray.array.
Any python methods which previously accepted numarray objects will accept numpy objects instead.
- The way solver options are defined for LinearPDE objects has been changed. There is now a SolverOptions object attached to the LinearPDE object which is handeling the options of solvers used to solve the PDE. The following changes apply:
- The setTolerance and setAbsoluteTolerance methods have been removed. Use now getSolverOptions().setTolerance
and getSolverOptions().setAbsoluteTolerance
- The setSolverPackage and setSolverMethod methods have been removed. Use now getSolverOptions().setPackage,
getSolverOptions().setSolverMethod and
getSolverOptions().setPreconditioner.
- The setSolverPackage and setSolverMethod methods have been removed. Use now getSolverOptions().setPackage,
getSolverOptions().setSolverMethod and
getSolverOptions().setPreconditioner.
- The static class variables defining packages, solvers and preconditioners have been removed and are now accessed via the corresponding static class variables in SolverOptions. For instance use SolverOptions.PCG instead of
LinearPDE.PCG to select the preconditioned conjugate gradient method.
- The getSolution takes now no argument. Use the corresponding
methods of the SolverOptions object returned by getSolverOptions()
to set values, e.g. use
getSolverOptions().setVerbosityOn() instead of argument verbose=True
and getSolverOptions().setIterMax(1000) instead of argument iter_max=1000
- The esys.pyvisi module from previous releases has been deprecated and will no longer be supported.
It is still present in the source code and can still be used if you compile esys.escript from source.
It will not be available in binary releases.
Its use is discouraged.
The documentation for esys.pyvisi can be found in Appendix B.
esys@esscc.uq.edu.au