gstlal-calibration:  Improved methods for doing noise subtraction multiple times using different witnesses in series.

fxtools/multichannel_datasource.py: fix issue with individual channels requested from INI files not being added. feature_synchronizer: fix issue where pipeline was crashing after some time due to not flushing messages from queue

datasource.py: add options in lvshm live source to set the assumed duration and block size of the shared memory partition

Makefiles are not configured to build itacac, and it will not work
without a patch for gstlal_snglinspiral that will be pushed yet

- test_commpn.py:  Seeking pipeline seems to work now.
- pcal-to-darm plotting script:  now works in new setup.
- gstlal_compute_strain: removed stray dq_epics variables

feature_extraction: tidy up options, fix hdf5 sink issues with reading ini files, extra tweaks in dag generation/Makefiles for online

gstlal-calibration:  Using Makefile to generate directories with filters when needed, but these directories are in .gitignore, so they will not be part of the gstlal repo.

gstlal-calibration:  removed git submodules with git-svn directories, which did not work the way we had hoped.

gstlal_compute_strain:  Split the PUM and UIM stages of the actuation and calculate kappa_PUM and kappa_UIM separately and apply them.

- reflects as-built changes included in the 1.5.1 packages.

feature_extractor: update synchronizer/sink to deal with variable sample rates, update Makefiles to process channel INI files seamlessly

(fixes bug where uploads with files from coinc_extractor wouldn't show
sngl_inspiral contents on gracedb, at least as recent as Feb 2018)

gstlal-calibration:  Added git submodules with calibration filters from the calibration SVN. The submodules were made using git-svn.

- unmaintained on master, auto* scripts becoming increasingly verbose about problems in the build scripts in that directory

- the detchar trigger generator stuff got moved to gstlal-burst but the
  packaging and build scripts got left with some stuff dangling here and
  there.  this patch comments out the relevant lines rather than removing
  them so that it's easier to re-instate libgstlalugly.so later if needed

gstlal_feature_synchronizer: fix issues where synchronizer was not combining the right subsets in some cases, add steps to ensure that buffers do not get processed more than once

gstlal_compute_strain:  pipeline searches for filters file given only file name in config file.  Order of preference for which filters to use is:
1) Filters from the calibration SVN (that is, those found in /ligo/svncommon/CalSVN/aligocalibration/trunk/Runs/*/GDSFilters/)
2) Filters in the user's home directory in a subdirectory called GDSFilters
3) Filters in the user's home directory in any subdirectory

- tests/check_calibration:  Caitlin Rose's Automated testing scripts and Makefiles for doing calibration tests.
- gstlal_compute_strain:  Don't demodulate the pcal channel again to compute SRC detuning if re-using the ~30 Hz pcal line
- lal_transferfunction:  1. Bug fixes for median of transfer functions.  2. Option to compute first transfer function asap at start of stream.  3. Use GPS time to decide when to update transfer functions so that output becomes independent of start time asap.
- tests/check_calibration:  Added plotting script and Makefile example for plotting time series of \DeltaL / pcal at calibration line frequencies.
- config_files:  Example config file for offline calibration with noise subtraction.

- increase minimum required framecpp to the version currently in production

- update to newer autotools scripts

- also somebody's been modifying this without putting "\" characters into
  lines.  I guess it worked ... but ... seriously?

  https://www.youtube.com/watch?v=FClGhto1vIg

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

multichannel_datasource.py: tweak settings and reading of channel INI files to make them much easier to use

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

feature_extractor.py: allow buffer in queue to be smaller than 1 sec to reduce overall latency, fix utils.floor_div to allow non-integer floor division

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

fxtools/feature_extractor.py: fix edge cases where wrong durations were being applied to hdf5 files, rename some methods to be more informative