Unit Tests

[StephMcCallum]

Unit tests are needed to check:

Periodic boundary conditions:

• particle outside of first periodic box
• non-cubic boxes
• long bonds

Random walk:

• uniform thetas and phis
• all particles are given a position
• all bonded particles are a bond length away

DPD frame instance:

• A particles are added to the frame
• positions from random walk are added to the frame
• bonds, angles and dihedrals are added to the frame

DPD simulation:

• parameters match input from kwargs
• running in a loop, checking simulation end criteria
• random number seeds

[erjank]

Seems like both a unit test and an overall sanity check would include using freud to calculate the RDF of our final DPD frame and report back the first few pair counts. E.g.:

"We have no nonbonded pairs for r<=0.85, and 15 for 0.85<r<=0.95, and 50 for 0.95<r<=1.05"
or
[dr=0.1,[0.9,1.0],[15,50]]

[erjank]

Implementing non-cubic boxes should be a separate issue after 1.0, but yes, checking for non-cubic boxes makes sense.

[erjank]

Unit test for uniform thetas and phis seems maybe overkill, but nice to have if it's not getting in the way.