Update Pastro Gstlal citation

Hi @brandon.piotrzkowski,

The GstLAL team is working on putting together a methods paper for Pastro summarizing the improvements implemented in O4. We are aiming to put it out on arxiv in a couple of weeks. Would it be possible to include the citation for it at that time? Thank you!

made the issue visible to everyone

changed the description

Note, I recently updated the gstlal citation for the pipelines here: #87 (closed)

We currently don't cite p_astro methods for individual pipelines, so there is no precedent. Although in the future the right thing to will be to have citations for every pipeline's p_astro and put this next to the p_astro classification, for now we could just add this to the existing gstlal citations. Let me know when this paper is out so I can add it.

It's not clear to me that GCN circulars are the right place to cite p-astro papers, especially if those papers are cited in the pipeline papers, like I believe this one will be in the new gstlal methods paper. If we cite p-astro implementations, then don't we also need to cite e.g. em-bright and template bank papers?

@cody.messick, @brandon.piotrzkowski I think em-bright is already cited: line 259, unless I am missing something. Also, GstLAL's citation contained the likelihood ratio methods paper in addition to the pipeline/performance paper despite the former being cited by the latter. Since the pastro methods paper will contain exact numbers corresponding to the rates used in the calculation as well as the template weights which determine the classification into BNS and NSBH etc., I thought it might be useful to have this paper as part of the citation provided of course it's allowed. Note that these numbers and pieces of information are not available in the main GstLAL performance paper. I just thought a citation to the pastro paper which contains these numbers and other relevant pieces of information, might be helpful to people who will be "using" the pastros computed from those numbers.

Bayestar also runs on all events :) I don't know why em-bright is singled out as drawing a parallel.

Yeah I think the key difference is that there are numerous instances of p_astro, one for each pipeline, that we would need to also cite to be consistent and move these properly the classification section. I'm not sure whether pipelines have separate papers for their p_astro but regardless this would require some coordination with pipeline people.

But besides the point, I can see how this could potentially become harder to maintain with all of the logic to draw citations based on which pipeline's p-astro it is. (well I don't maintain it, but can sympathize with @brandon.piotrzkowski :) ) Also the code needing constant synced releases of ligo-followup-advocate and gwcelery which adds extra work. A long term and easier solution would be to simply add all references to the userguide.

Maintenance is clearly a problem now as we are getting demands from the RRT to make new release just to get the citations up-to-date.

It might seem like a stretch but I suggest dropping the citations entirely, with a link to the Userguide which will list all citations. It keeps our already long circulars not becoming even longer. This has been a problem in every observing run. This might be something to bring up at the LL and Ops call this week?

I don't think a change like that would be for the release set for this week, but maybe something to discuss for a future release.

We could discuss this on calls this week while I prepare a release fixing other issues.

changed milestone to %O4a mid-run release

changed title from Update Gstlal citation to Update Pastro Gstlal citation

The paper has been uploaded to arxiv: https://arxiv.org/abs/2306.07190