gstlal_compute_strain:  add filtering for nonsens subtraction channel in case re-timestamping or high-pass filtering is needed

 - Modified calibration state vector for O4
 - Added new line amplitude channels for calibration line subtraction
 - Added NonSENS subtraction channel
 - removed dewhitening options
 - removed kappa_PU stuff
 - removed some (maybe all?) kappa_A stuff
 - removed old version of calibration line subtraction

calibration_parts.py:  More efficient and flexible method of converting array properties to GstValueArrays

added elements lal_matmix and lal_tdepfirfilt to remove dependence on GValueArrays without having to change gstlal

pcal2darm_timeseries.py:  Added a plot showing R(f) thermalization as numerous R(f)'s on the same plot

calibration_parts.py:  Simplified some functions with a mkgeneric() that checks for array properties

lal_complexfirbank:  No need for GstValueArray read/write functions, as these are in the GstValueArray branch of gstlal

Numerous fixes for compatibility with new pyGObject, especially use of GstValueArray.  The related (dependency) branch in the gstlal repository is called GstValueArray.

Add a flake8 CI job

Closes #15

See merge request Calibration/gstlal-calibration!5

currently just ignores tab indentation (W191)

this job should 'never' fail, but will present results in the code quality
widget on the merge request UI, see

https://git.ligo.org/help/ci/testing/code_quality.html#code-quality-widget

all dependencies

Adding gstlal_correct_strain pipeline

See merge request Calibration/gstlal-calibration!3

Remove duplicate usr/bin entry from Debian install files

Closes #4

See merge request Calibration/gstlal-calibration!2