Some features exist for `TempoPulsar` but not for `PintPulsar` objects.

nanograv / enterprise

ENTERPRISE (Enhanced Numerical Toolbox Enabling a Robust PulsaR Inference SuitE) is a pulsar timing analysis code, aimed at noise analysis, gravitational-wave searches, and timing model analysis.

https://enterprise.readthedocs.io

MIT License

65 stars 67 forks source link

Some features exist for `TempoPulsar` but not for `PintPulsar` objects. #303

Open AaronDJohnson opened 3 years ago

AaronDJohnson commented 3 years ago

I'm writing this issue as a request to add certain features to the PintPulsar class which already exist for the TempoPulsar class.

~~Add ability to get DMX values from PintPulsar class~~
Load PintPulsar into shared memory to reduce total memory required when using PTMCMCSampler
Add missing planet vectors (Mercury, Venus, Mars, Pluto) to PintPulsar

Are there other differences that I have missed?

vallis commented 2 years ago

Features 1 and 2 can be put in... 1 is easy if you know where to find the information in PINT (I don't, but could look). 2 is almost there, it just needs restructuring _flags into a numpy record array.

As for 3, I think that PINT does not provide those planets, so we should open an issue on the PINT repository. Right @Hazboun6 ?

scottransom commented 2 years ago

Why are those planets necessary? We can totally add them, but Mercury and Pluto?!? And I suspect that even Venus and Mars are way more precise than we need, no?

vallis commented 2 years ago

They're not necessary. We're just getting carried away by feature parity... they are not a priority.

vallis commented 2 years ago

Also, I just checked and it looks like .dmx (as needed by wideband data) is already in PintPulsar. @AaronDJohnson, is there something specific you can't find?

AaronDJohnson commented 2 years ago

Also, I just checked and it looks like .dmx (as needed by wideband data) is already in PintPulsar. @AaronDJohnson, is there something specific you can't find?

This one was from a discussion with someone the other day. It must have been added recently, and I should have checked!

Hazboun6 commented 2 years ago

@scottransom and @vallis I agree that most of those are probably not needed. And definitely don't need to be a priority.

I will say that the PPTA coded up a version of BayesEphem with perturbations to Mars' orbit when they ran their most recent GWB analyses. They have excess power at a frequency near to the inverse period of Mars' orbit. So they did use the Mars information in the Tempo2Pulsar objects.