Prevent Redundant Definition

[pulsipher]

This closes #146. It also allows us to use macro defined derivatives repeatedly without erroring unlike before.

With v0.4.0:

function test()
       m=InfiniteModel()
       @infinite_parameter(m, t in [0, 10], derivative_method = OrthogonalCollocation(4), num_supports = 100)
       @infinite_parameter(m, x[1:2] in [0, 1], num_supports = 20)
       @variable(m, y >= 0, Infinite(t))
       @variable(m, q, Infinite(t, x))
       @objective(m, Min, integral(y + integral(q^2, x), t))
       @constraint(m, deriv(y, t) == 2y + 2)
       @constraint(m, deriv(q, t) == q + y)
       @constraint(m, y * q - 2 <= 2)
       build_optimizer_model!(m)
end
@time test()

1.228544 seconds (6.01 M allocations: 388.745 MiB, 7.29% gc time)

With this branch:

@time test()

0.939251 seconds (4.64 M allocations: 311.325 MiB, 11.42% gc time)

Hence, we have a significant decrease in the amount of memory and allocations. However, there isn't really a speed up since redundant definition wasn't a rate determining step.

[codecov]

Codecov Report

Merging #153 (6cab6a6) into master (a275ada) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff            @@
##           master     #153    +/-   ##
========================================
  Coverage   99.96%   99.96%            
========================================
  Files          32       32            
  Lines        6185     6286   +101     
========================================
+ Hits         6183     6284   +101     
  Misses          2        2            

Impacted Files	Coverage Δ
src/TranscriptionOpt/model.jl	100.00% <ø> (ø)
src/TranscriptionOpt/transcribe.jl	99.69% <ø> (-0.01%)	⬇️
src/TranscriptionOpt/measures.jl	100.00% <100.00%> (ø)
src/datatypes.jl	100.00% <100.00%> (ø)
src/derivative_evaluations.jl	100.00% <100.00%> (ø)
src/derivatives.jl	100.00% <100.00%> (ø)
src/finite_variables.jl	100.00% <100.00%> (ø)
src/infinite_variables.jl	100.00% <100.00%> (ø)
src/measure_expansions.jl	100.00% <100.00%> (ø)
src/point_variables.jl	100.00% <100.00%> (ø)
... and 4 more

---

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 a275ada...6cab6a6. Read the comment docs.