- this program crashes if the injection file is empty.  why is it empty?  it just is, OK.  sheesh.
- this patch reworks the PSD --> GPS time mapping code so that it doesn't crash when there are no injections.  a significant amount of simplication is possible, too, and so this patch also does a buch of cleanups.
- note that this program is borken for anything other than "H1", "L1", "V1".  some generalizations are made, here, to simplify the expansion to different detector networks in the future, but I don't actually fix the lack of, e.g., KAGRA support at this time

- don't hard-code instrument list, get from PSD files
- use plotutils to convert instrument name to colour (and color)

dropped in online ranking statistic code

- Param elements now do support None, but not via the .from_pyvalue() method, yet, only via the .build() method.

gstlal_feature_extractor: ensure that filenames of hdf5 files containing features have accurate gps range in their names by using the feature_start(end)_time options, if specified. changed named of trigger_start(end)_time from trigger -> feature. removed unnecessary typecasting for gps times

gstlal_feature_extractor + idq_utils.py: modified method of saving features as hdf5 so that max_snr option produces regularly sampled feature rows now

gstlal_compute_strain:  Added calibration line subtraction back in. There's also now a prototype function for subtraction of 60 Hz lines and harmonics, and for noise subtractions using witness channels.

gstlal_peakfinder.c: fixed issue in gstlal_peak_max_over_channels where peak_max was not stored correctly, producing nonsensical results

- ignore instruments that haven't produced any triggers at all, because we
  wouldn't be worried about ranking those candidates anyway, but require at
  least one instrument to have produced triggers

doc Makefile: remove ignoring of __init__.py in gstlal/python, appears to trigger an AssertionError which is a known bug within sphinx

doc/Makefile: exclude gstlal/python/__init__.py from sphinx autodoc generation as this causes issues with building gstlal python documentation within gstlal proper

gstlal_feature_extractor: add dot graph for documentation, added separate groups for command line options, move imports around for aesthetics, change comment blocks (e.g. main) for documentation

Revert 9ac1b181

See merge request lscsoft/gstlal!1

This reverts commit 9ac1b181