- also reduce the line width of both plots to cause the PSD plot to hide fewer features

- introduces summarize_coinc_xmldoc()

- axes_plot_psds() generates a PSD plot on an axes object.
- plot_psds() is now only a wrapper that creates a figure object and uses axes_plot_psds() to populate its one set of axes with the PSD plot.
- this allows the PSD plot to be added to other, multi-axes, figures in the future.

- change the range of frequencies over which the spectrum min/max is measured to [10, 900] from [10, 1000]
- choose the lower bound of the y-axis by dividing the spectrum min by 3 before taking the log and the floor to provide some extra padding below the spectrum
- increase the number of y axis ticks and add minor ticks to the y axis

matplotlib plotting error (axes.plot cannot take markersize = None, at
least not in version 1.4.2, the current CIT version)

- temporary hack, remove after O2

- to an error message

- forgot one spot in plotpsd.py
- refs #5603

idq_multirate_datasource.py: fixed normalization in downsampler, fixed dc offset that appears in certain channels

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

The SNR time series epoch has to take into account the original
template's time shift.

- the version inherited from dict returns a dict instead of an
  HorizonHistories instance, and makes references to the values in the
  original object so that modifications affect both the original and the
  copy.
- this override returns the correct type, with deepcopy()s of the values

- introduce .functional_integral() methods to both NearestLeafTree and
  HorizonHistories classes
- reimplement .weighted_mean() as a wrapper of .functional_integral().
  this results in a performance penalty since it must invoke
  .functional_integral() twice but avoids the code duplication that would
  result otherwise
- allow 0/0 in .weighted_mean().  if the numerator of the mean is
  identically 0 and the weighted denominator has also evaluated to 0 then
  return 0 instead of reporting a divide-by-0 error.  this is done
  presuming that this is caused by the weight function being defined in a
  way that causes it to go to 0 when the data being averaged goes to 0 and
  that if a proper limit of the ratio was evaluated it would be found to be
  0.