merge DiffusionProcess1D into DiffusionProcess

[tlestang]

Edit 01/07/2020 This PR grew larger than just extending DiffusionProcess.update. It ports all methods from the specialised 1d class DiffusionProcess1D to the general class DiffusionProcess, using the @method1d decorator to restrict some of them to process in 1D

---

At the moment it is assumed that if DiffusionProcess.update is called from a process with dimension=1, then the state vector is an array of size 1. This adds a conditial in DiffusionProcess.update to make update return a float for `1d processes.

I chose not the check the dimesion with something like

if len(xn) != self.dimension:
  raise ValueError("State vector xn has dimension {}, expected {}".format(len(xn), self.dimension))

Reasons:

• If update is intended to be mostly called internally, it's not required
• It may hurt performance if update is called repeatedly
• If xn is float then len(xn) does not exist, so must first use extra np.atleast1d which adds complexity.

[codecov-commenter]

Codecov Report

Merging #20 into dev will decrease coverage by 0.27%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##              dev      #20      +/-   ##
==========================================
- Coverage   77.84%   77.56%   -0.28%     
==========================================
  Files          14       14              
  Lines        1124     1110      -14     
==========================================
- Hits          875      861      -14     
  Misses        249      249              

Impacted Files	Coverage Δ
stochrare/dynamics/diffusion.py	85.40% <100.00%> (-0.39%)	:arrow_down:
stochrare/utils.py	100.00% <0.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 67e97d9...244af0f. Read the comment docs.