…bdot

…bdot

…bdot

…bdot

…bdot

… cbd_abdot

…bdot

…onics) to GW_Discrete instance

…softening instead of hardening-only)

Use time (lookback time) as independent variable instead of binary separation to allow for binary 'softening' (i.e. da/dt > 0).
Integrate each binary, one at a time, from initial conditions until either coalescence or redshift zero.  Evolution arrays thus
are no longer always the same size.  Currently evolution arrays are 1D objects, flattened for all binaries, and the appropriate
indices are stored (Evolution._last_index).  This *seems* to be working (first full implementation).  Lots more changes will be
needed to allow for interpolation, universe-realization, etc.  Its also extremely slow.  There are numerous obvious problematic
places, e.g. interpolation of CBD_Torques and Dynamical Friction halo relations.  Cython could also speed things up a lot, but
the issue will be external function calls.

Really, this should be a second Evolution class, so that both implementations are preserved for the moment... I'll add the old
one back in later.

…pulation structure

…rking.

BUG: did catch an issue where the first/last target frequencies wouldnt always be found.
Disabled the resetting of the target-frequency index when changing directions.  Tests pass even without it... so hopefully okay?

…side it, with the new-style of integration.

Also get the interplation for the new version working.

…ff breaks

Add simple implementations of some utility functions so that they can be called from cython without much overhead.  Perhaps should change these to be cython implementations called from python (or cpdef functions).

…take_next_step

* origin/cbd_abdot:
  new tests and changes in notebooks
  m1m2 order for both evo classes, output change in velocity_orbital
  hardening compatible with new/old evo
  evolution compatible with old/new evo, fixed mass switching issue in take_next_step
  make accretion.py compatible with both new and old evo
  initialize self.mdot=0 whether or not we use accretion, otherwise stuff breaks
  no need to set mdot=0 at >1.e5pc separations
  notebooks
  dfdt with accretion

…grav/holodeck into cbd_abdot

…eters in evolution init

* cbd_abdot:
  move interp and pkl.load to init
  cfl condition for eccentricity evolution, add nsteps and cfl as parameters in evolution init
  add ebeq function and move interpolation and pkl loading to init
  changes in notebook

… cbd_abdot

… cbd_abdot

… script for New_Evolution

… cbd_abdot

… cbd_abdot