- add alternate, faster, implementation using pandas (fall back to current implementation if pandas is not available)

 - gstlal_inspiral_pipe and gstlal_svd_bank take as option
--sample-rate
 - the waveforms that are generated with a lower fhigh
are shifted forward such that all waveforms in a given
sub bank have a common amount of early warning
 - this has the effect that the actual fhigh of waveforms
is less than requested in a given sub bank, since the
waveforms are shifted according to the waveform that takes
longest to go from fhigh to ISCO

Signed-off-by: Surabhi Sachdev <surabhi.sachdev@ligo.org>

- switch to SEOBNRv4 waveform model for accuracy at higher masses
- introduce mechanism for pre-computing waveform model for speed
- refs #5603

cbc_template_fir.py: Introduced a fir whitener with low latency and if-statement as a switch between old and new whiteners

- in generate_template() we used to have to convert the swig fseries object
  to a pylal fseries object.  we don't have to do that anymore but the
  conversion code was still there.  remove

This reverts commit 5f9d2a3f.

and fixes it by making adjustments to pipeio.py, reference_psd.py, inspiral.py

- make add_quadrature_phase a static method

This reverts commit 3111aaf6.

- replace with lal's swig-wrapped LIGOTimeGPS

- glue and pylal both all export the .end property so this
  modification creates compatibility with both
- I hope to update the lal swig bindings soon as well
- refs #2288

- port code to functions in lalsimuation.  also change the padding formula
  from 1.1 * tchirp + 1 to 1.1 * tchirp + 3/flow.
- even with the original padding formula the results are slightly
  different.  e.g., with the ER7 "em bright" bank and an f low of 10 Hz,
  the old implementation determined the padding to be 175.7 s with a padded
  f flow of 9.647 Hz, while the same formula implemented with the
  lalsimulation functions determines the padding to be 179.8 s with a
  padded f low of 9.583 Hz.
- this also adds some input checking to avoid getting weird error messages
  from this code
- refs #2179

A new table contains a link to a diff between the current HEAD of master
and the git hash at the time of review.

cbc_template_fir.py: added review documentation and addressed action item to remove two dead functions

2) Add IMRPhenomC, SEOBNR_v2_ROM_DoubleSpin in
   the list of gstlal IMR approximants and remove
   IMRPhenomB, since it follows a different epoch convention.

add the duration over which the psd is calculated to the template length for psd ringing instead of 32s

- there was a mysterious off-by-one hack included in the firbank latency
  in lloidparts that had been found to be required in order to get the
  impulse response from the different template fragments to line up
  properly.  this patch doesn't really explain the mystery, but moves
  the logic from the firbank into the template slicing logic so that it
  all lives in one place, and corrects an overall off-by-one latency
  error while at it.