python reader for native file format

[danieljprice]

this is posted here as a reference, currently the supported solution for this is to use libphantom to call the read_dump routine directly, or to convert to another format (e.g. using splash or phantom2hdf5)

A better solution would be to have a pure python reader for the native format. A preliminary version of this was kindly contributed by Lionel Siess and Ward Homan, based on work by Stéven Toupin. Just needs to be cleaned up and put into a separate repository (i.e. as a standalone python package)

This issue is just a reminder that this needs to be done, and to flag that this exists for anyone trying to do this.

[lsiess]

The file scripts/readPhantomDump.py is now available in the master branch. It works most of the time. Note that handling hdf5 files has never been tested. Here is a simple example on how to use it :

from readPhantomDump import *
from pylab import *

fdump='/my/path/to/run_00041'
data3D = read_dump(fdump)
fpar='/my/path/to/run.in'
data_in=read_infile(fpar)

tmax=data_in['tmax']
npart = data3D['part_numbers']['npartoftype'][0]
udist, umass, utime = [data3D['units'][u] for u in 'udist', 'umass', 'utime']
hfact = data3D['quantities']['hfact']
mpart = data3D['quantities']['massoftype'][0]*umass
x, y, z, vx, vy, vz, h, u = [array(data3D['blocks'][0]['data'][c],dtype=float)*conv for c, conv in [('x', udist), ('y', udist), ('z', udist), ('vx', udist/utime), ('vy', udist/utime), ('vz', udist/utime), ('h', udist), ('u', udist**2/utime**2)]]
r = sqrt(x**2+y**2+z**2)
v = sqrt(vx**2+vy**2+vz**2)
rho = mpart * (hfact/h)**3

# Plot velocity
figure(1)
clf()
plot(r,v/1.e5, '.', color='b', markersize=1.592, rasterized=True, label='3D')
tight_layout()
show()

Cheers Lionel

[dmentipl]

Hi @lsiess, just to let you know: for reading Phantom HDF5 files there's also Plonk. If you use it and have any issues, you're welcome to raise over at the Plonk issue tracker.