Refactor NRSur7dq2 and fix for waveform review
-There are now two approximants: NRSur7dq2_HM uses all modes available (ell <= 4) NRSur7dq2_L2only uses only the (2, +/-2) modes in the co-orbital frame -Use lalsimulation lalinference_o2 conventions in ChooseTDWaveform As there is not a consistent convention, we follow SEOBNRv3. -Fix various issues with GPS time, fmin, fref -Add StartFrequency function giving smallest allowed fmin and fref -Add documentation -Fix small numeric error in waveform evaluation