convolve with crossing fibres (no dispersion)

Each function takes as input:

• compartment
• all acquisition parameters required by compartment (should include bvec)
• fODF parameters
  • for non-SH: list of tuple (volume fraction, fibre orientation / fibre orientation + kappa) for example: crossing = [(0.3, 'delta', [1., 0., 0.]), (0.7, 'watson', ([0, 1, 0], 4.)]

def convolve_crossing(compartment, acquistion, crossing):
   for volume_fraction, dtype, param in crossing:
       func = {'delta': convolve_delta}[dtype]
       signal += volume_fraction * func(compartment, acquisition, params)

Output is the signal as 1D array

General integral: S(g, b) = \int fODF(\nu) RF(b, g \dot \nu) d \nu

In our specific case the fODF is a bunch of delta functions, so: S(g, b) = \sum_i f_i RF(b, g \dot \nu_i)