Updated algorithm that fits splines into beads to extend farther into intrahelically connected segments, and to use linear interpolation for quaternions, resulting in better all-atom models

Updated bead model potentials to work with antiparallel DNA; also made terminal_crossovers bond shorter at 1.2 nm (vs 1.85)

Added type_ argument to SingleStrandedSegment.add_crossover() for compatibility with DoubleStrandedSegment

Updated imports to be compatible with new package structure; updated paths to resources to use new get_resource_path() function

Made bond potentials longer ranged (120 vs 50 nm), and removed unneeded 'continue' statement that was attempting to prevent double-counting when placing initial beads

Fixed spring constant for dihedral that spans crossovers, keeping helices parallel (note this differs from prototyped web service where 90 degree angle potentials were used)

Changed dna bead type clustering to have higher resolution

Streamlined some small parts of the code

Moved run_simulate_protocols outside of cadnano_segments since more modular; increased density of coarsest DNA model

Made intrahelical connections remove any existing connections; important for intrahelical "crossovers"

Added function for setting splines; converted an assertion into a warning when beads are very close; fiex crash when helices lack crossovers

Improved cadnano segment protocol; using 1.03 instead of 1 for max beads, resulting in more 1-bp beads

Now clustering nucleotide types rather than rounding; gives more accurate number of nucleotides in a double helix