Wragged output methods

[ben18785]

Implements SubProblem and ProblemCollection methods which collectively allow:

• problems to be constructed where different outputs are measured at different times
• separate subproblems to be created for different subsets of outputs (and, accordingly, for inference, given different score functions or log-likelihoods)
• caching of simulation outputs between evaluations for the same parameter set

[codecov[bot]]

Codecov Report

Merging #1405 (d68edbb) into master (d64eb05) will not change coverage. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##            master     #1405    +/-   ##
==========================================
  Coverage   100.00%   100.00%            
==========================================
  Files           87        87            
  Lines         8889      9018   +129     
==========================================
+ Hits          8889      9018   +129     

Impacted Files	Coverage Δ
pints/__init__.py	100.00% <100.00%> (ø)
pints/_core.py	100.00% <100.00%> (ø)
pints/_optimisers/_pso.py	100.00% <0.00%> (ø)
pints/_optimisers/_snes.py	100.00% <0.00%> (ø)
pints/_optimisers/_xnes.py	100.00% <0.00%> (ø)
pints/_optimisers/_cmaes.py	100.00% <0.00%> (ø)
pints/_optimisers/__init__.py	100.00% <0.00%> (ø)
pints/_optimisers/_cmaes_bare.py	100.00% <0.00%> (ø)
pints/_optimisers/_nelder_mead.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 d64eb05...d68edbb. Read the comment docs.

[ben18785]

@MichaelClerx would you be able to take a look here if you get a moment?

[MichaelClerx]

I'll try to look before too long. Are there no PKPD modellers who could have a look in the meantime?

[ben18785]

Thanks! @martinjrobins -- if you had a spare few mins, it would be good to see what you thought about this.