----------------------------------------------
                  C.O.W.

parallel analysis tools for the Cube of Wonder

Author: Jonathan Zrake, NYU CCPP
----------------------------------------------

The functions provided in this library are written to do analysis of
hydrodynamic and magnetohydrodynamic turbulence, especially on the unigrid
periodic cube, A.K.A. the swirly cube or Cube of Wonder.

Everything here is written in C and Python. Optional libraries are MPI, HDF5,
and FFTW.


Features planned or implemented
-------------------------------
[+] ... finished
[-] ... started
[ ] ... planned
-------------------------------

[+] Support for 1d, 2d, and 3d data domains with as many named data fields as
    you want

[+] Domain decomposition: with arbitrary depth guard cell synchronization

[+] Parallel data I/O routines: read/write with or without MPI collective HDF5

[+] Histograms: in 1d and 2d, linear/logarithmic spacing with weights, HDF5
    writable

[+] Transforms (including stencils) via C callback functions

[+] Vector field derivative operations: div, grad, curl, etc.

[+] Parallel 3d FFT functions: allows arbitrary data layout, no need to keep 3d
    data in 2d slabs

[ ] 2d parallel FFT's

[-] Serial versions (for use without MPI) of 1d, 2d, and 3d FFT's

[+] Helmholtz decomposition of vector fields

[+] Sampling of data points on remote processors: by global index or
    interpolated to physical locations

[ ] Support for domains with complex-valued data

[-] Two-point correlations, structure functions

[ ] Bi-spectra

[ ] Tracer particles

[ ] Vector field streamline integration

[ ] Low/high pass filtering; course-graining

[+] Python wrappers