Add --polycos option for phase calculation

[sguillot]

The new option --polycos in the PINT photonphase script significantly improves the calculation time for large event files (nanograv/PINT#1199).

We should consider adding this option to psrpipe.py and merge.py.

[paulray]

Indeed! As soon as that is tested and merged, we should take advantage of it!

[paulray]

However, for NICER data there is a significant optimization needed. It shouldn't be hard but Sasha doesn't have time to work on it.
nanograv/PINT#1209

[sguillot]

@paulray, what is the status of the use of --polycos in PINT ?

[paulray]

It should be working, but it requires barycentered photons.
Please try it out!

[sguillot]

This could be implemented at the same time as #28, i.e., psrpipe could call barycorr if needed.

The question is which coordinates to use for the barycentric correction:

1. from the event file header ?
2. user-provided coordinates ?
3. the PAR file ?

Thoughts?

[paulray]

Yes it could barycenter the data. That would create a new evt file, which would have the phases, but actually the PULSE_PHASE column could even be copied back to the topocentric event file. The barycentering should use the .par file coordinates, but there are options here: (1) use RAJ,DECJ or (2) use RAJ,DECJ accounting for PMRA and PMDEC or (3) use something other than barycorr to do the barycentering. It will be important to match the ephemeris used for barycorr with the one specified in the par file.