Pyclaw is a Python-based solver for hyperbolic PDEs that includes the algorithms of Clawpack and SharpClaw. It has been designed with easy extensibility, performance, and exploration in mind. PyClaw is also the basis for PetClaw, a scalable parallel implementation of Clawpack using PETSc. At present, PyClaw also includes a python toolkit containing various other routines to help users use Clawpack.
You can get the latest development version of PyClaw from http://github.com/clawpack/pyclaw.
Full documentation can be found at http://numerics.kaust.edu.sa/.
PetClaw is a parallel version of Clawpack. The high-level code is written in Python using numpy and based on the PyClaw package. The low-level kernels are written in Fortran and based on Clawpack. The parallelism is implemented using PETSc and petsc4py.