- these are being used temporarily for debugging the online pipeline;  this patch is my current state so Chad can pick it up and see if he can make progress

- --likelihood-file is renamed to --ranking-stat-output to make its function more clear
- --likelihood-file-cache is renamed to --ranking-stat-output-cache to retain the symmetry
- --reference-likelihood-url is renamed to --ranking-stat-input to make its function more clear

- to --ranking-stat-pdf to more correctly describe what this command line option does.
- also switch to using the load_url() function, instead of load_filename(), to anticipate getting this data from a non-disk location in O3.

Kipp Cannon authored 6 years ago and Kipp Cannon committed 6 years ago

- correct documentation and add missing documentation for command line options
- --help message is reorganized into conceptual groups
- rework input sanity/safety checking and associated error messages
- make --likelihood-url optional, no file will be written if not supplied, and trigger histogram collection will be skipped for performance improvements
- teak gstlal_inspiral_pipe to not set --likelihood-url options on injection jobs

- closes out two FIXMEs

bin/gstlal_inspiral_create_dt_dphi_snr_ratio_pdfs: program to figure out what probability you have of forming coincs with certain parameters

gstlal_inspiral_create_p_of_ifos_given_horizon: program to figure out what probability you have of forming coincs in an instrument network

-the code needs to know how many input samples are needed to drain enough of the history for there to be nothing but 0s remaining.  to determine the amount of non-zero history, instead of counting contiguous non-zero samples from the start of the history count contiguous zero samples from the end and subtract from the total.

gstlal-calibration:  adjust gstreamer plugin .so filenames on gstreamer 1.14+ to work around changes to name mangling used by the plugin scanner

Kipp Cannon authored 6 years ago and Kipp Cannon committed 6 years ago

- when collecting trigger histograms from scratch, initialize the numerator with the signal model so that the resulting ranking statistic file is closer to being in a usable state as-is, without having to mix in the output of gstlal_inspiral_create_prior_diststats
- this has been made possible by the recent change to the numerator's .__iadd__() method that causes overall histogram count to be preserved across marginalization operations
- teach gstlal_inspiral_pipe to disable numerator initialization in the create_prior_diststats jobs since it is not longer required

Kipp Cannon authored 6 years ago and Kipp Cannon committed 6 years ago

- the count is meaningless, it is only used to set the scale of the density estimation kernel, but for that reason it should not change as the result of addition operations
- in the current pipeline care is taken to never add more than a single copy of the numerator into any ranking statistic in order to control the total count
- this patch causes .__iadd__() to preserve the total count, allowing the analysis pipeline plumbing to be simplified since there is no reason to fear adding multiple copies of the numerator together

Koh Ueno authored 6 years ago and Kipp Cannon committed 6 years ago

- adjust gstreamer plugin .so filenames on gstreamer 1.14+ to work
around changes to name mangling used by the plugin scanner

updated offline feature extractor Makefile to have configurable channel list as well as updated job parameters for submitting Condor jobs efficiently

- allow parameter to be controlled from the command line.

- options that can be given multiple times and that have default values were appending the values from the command line to the defaults instead of overriding the defaults with the user-supplied values.

- enable 1 and 2 by default

- when initializing an SNR PDF cache with new entries, a math domain error
  can occur when an instrument combination is impossible.  this patch fixes
  it
- it's very odd that this code worked in the past.  Python must be changing
  its behaviour.  commit 29a7482f contains
  both the original version of this code and the data file it successfully
  generated at the time ... so ... it must have worked.  I don't know what
  else to say, but I'm not crazy.

multichannel_datasource.py: added option to force splitting of channels into subsets with an equal number of channels for serial processing

multichannel_datasource.py: added a 4s max buffer duration reblock element after demux when reading in frames to reduce memory footprint substantially