gstlal_compute_strain:  Only use UIM line for computing kappa_U. Added another coherence uncertainty bit the the calib state vector.

- rewrite streamthinca as a direct interface to the new streaming
coincidence engine in snglcoinc.py instead of being a wrapper around the
offline coincidence engine in thinca.py

Update .gitlab-ci.yml: switched to system-build of gstlal deps in el7 container. general clean up. this switch involves migration of to gstreamer 1.14

pipeparts: work around for broken bindings override for elem.get_request_pad(). FIXME this is a dubious workaround that should be addressed more properly and eventually reverted when the libraries on the reference OS are updated and it is no longer a bug.

added --disable-massmodel option to turn off installation of lnP_template_signal_BBH_logm_reweighted_mchirp.hdf5. Made madification to .gitlab-ci.yml

Adding in tests for gstlal-calibraiton-1.2.1 release, fixing a few of the test scripts, and updating gstlal_compute_strain to use pipeparts mkchannelmux function

- "if coinc_event_index" is inside "if last_coincs", which already is testing that there are coincs in the index.

This reverts commit 9159362d.

- this is to simplify changes that will be required as part of the coincidence code rewrite.  want to reduce references to last_coincs

- .from_pyvalue() now supports None

gstlal_inspiral_create_prior_diststats: remove chads mass mode file in favor of Heather's population model file

inspiral_lr: remove chads mass mode file in favor of Heather's population model file.  Additionally make sure that the iadd method can handle situations where not all objects have a population model specified, but force them to be the same if they do

gstlal-calibration/tests/check_calibration:  Improvements to pcal2darm_timeseries.py and plot_transfer_function.py