bin/gstlal_inspiral_create_prior_diststats: just call the add_foreground_snrchi_prior in place of the add_foreground_prior method which was not doing exactly what we wanted anyway

far.py: remove adding snr/chisq prior during add_foreground_prior() and change the name of that method to be what it actually does now

gstlal_inspiral_calc_rank_pdfs: remove ability to specify foreground or background prior and change the method of add_foreground_prior to be what it actually does

gstlal_inspiral_calc_likelihood: remove ability to specify foreground or background prior and change the method of add_foreground_prior to be what it actually does

- it seems all the work that has gone into making the sampling faster paid
  off without me noticing until now.  instead of taking 90 minutes, this
  now completes in 7.  so let's increase the default number of samples 10x
  and let it to run for a more reasonable length of time.

- modify RankingData to track the zero lank ranking statistic values explicitly,
- modify compute_far_from_snr_chisq_histograms to collect zero lag ranking
  stat values when doing its count-above-threshold pass
- while we're at it, switch the binning to ATanLogarithmicBins, and make
  other adjustments to move RankingData in the direction of becoming a
  subclass of CoincParamsDistributions

gstlal_inspiral_pipe: fix typo in previous commit regarding output location for the gstlal_inspiral_create_prior_diststats jobs

Kipp Cannon authored 10 years ago and Kipp Cannon committed 10 years ago

- need to change api of .get_snr_joint_pdf() so that we can avoid isntantiating a ProgressBar() when not needed

- if the loading of an input file fails, try processing the rest and then
  exit with an error code afterwards to reduce the time spent in rescue
  dags

- copy --synthesize-injections and --background-prior options from
  gstlal_inspiral_calc_likelihood

gstlal_inspiral_pipe: use calc_rank_pdfs to get final distributions and turn off sampling in gstlal_inspiral_calc_likelihood

This change makes a few changes that have been lingering. All of them are utlimately related to performance and robustness of the DAG.

1) Break up the last database merging step across mass bins into parallel chunks
2) Add a post script to check the file integrity of gstlal_inspiral output files
3) Add placeholder jobs for create_prior_diststats and calc_rank_pdfs to be able to parallelize the likelihood computation

description of a makefile which automates installations of released versions of
lalsuite and gstlal on vanilla machines while using optimized dependencies.
Also updated install direction and versions used for users wishing to do handle
install by hand

- switches to Jolien's new injection conditioner wrapper to generate clean
  inspiral injections.

- add framed=true to the allowed sink caps, and have element switch to pass-through when input is framed.  allows element to exist in pipelines that don't need it without breaking pipeline

to try to avoid a typecast

- instead of processing .sqlite files, process .xml format files

- reduces (but does not eliminate) risk of leaked memory