mhardcastle / pysynch Goto Github PK

Unless I'm missing something the zeta parameter to account for non-radiating particles is defined differently for the 'equipartition' and 'minimum_energy' methods in synchro.py . For the former it seems to be defined as 1/k where k is the ratio of energy density in radiating particles to the total energy in particles, while for the latter zeta=k. It would be helpful if this was consistent and also if zeta defined in the documentation.

I'm trying to measure batch properties (B, Pint) for samples of 100s to 1000s of objects in a loop that calls SynchSource and then normalises it to set the properties. This is running successfully for large numbers of sources and then hanging in a non-reproducible way on the s.normalize step -- repeated runs have it stopping at a different source, and whichever source it stops at I can successfully run individually by hand with the same parameters.

The code I'm running is /beegfs/lofar/jcroston/surveys/hizfeedback/runsynch.py and called as follows:
% python3 runsynch.py unresolved24_withenvsx_kcorr.fits test_out.fits

On my last two runs it hung at source 290 and 1443 respectively - this seems very random - but is not just an issue flushing the diagnostic output as I've left it overnight to see if it wakes up and it remains stuck.

It seems to work OK when updating "synchmodule.c" so that it looks like this:

//(void)Py_InitModule("synch", SynchMethods);
static struct PyModuleDef synch =
{
PyModuleDef_HEAD_INIT,
"synch", /* name of module /
"", / module documentation, may be NULL /
-1, / size of per-interpreter state of the module, or -1 if the module keeps state in global variables. */
SynchMethods
};

PyMODINIT_FUNC PyInit_synch(void)
{
// do some initialization

w1=gsl_integration_workspace_alloc(GSL_WSIZE);
w2=gsl_integration_workspace_alloc(GSL_WSIZE);
gsl_set_error_handler_off(); // to avoid abort
gsl_epsilon=1e-3;
gsl_quiet=1;

fx=calloc(xls,sizeof(double));
#ifdef DEBUG
printf("running makefx\n");
#endif
makefx();

// these numeric parameters and the spectral function to be used are
// changeable by the user: set some defaults

gmin=10;
gmax=100000;
set_minmax_globals();
power=2.2;

set_simple_powerlaw();

return PyModule_Create(&synch);
};