Update doc

[guimarqu]

Preview at https://atoptima.github.io/Coluna.jl/previews/PR564/

1. fix CapacitatedVehicleRouting demo utilization problem #498

• pricing callback example with Literate and KnapsackLib
• more information about the column generation algorithm
• more documentation about specify and multiplicity (2.3)

2. fix can Coluna decompose a general large scale MILP model?  #539

• more documentation about the axis, what it returns, how to use it
• warn that anonymous constraint and variables are not supported
• decomposition (specify and multiplicity) example with Literate + use of annotation

3. fix No primal primal bound? #545

• documentation about init primal and dual bounds
• documentation about artificial vars costs
• more information about algorithms
• Lara: behavior of SolveLpForm when the algorithm is only used to solve a linear problem

4. fix Decompose twice at the same level. #560

• documentation about dantzig-wolfe decomposition & reformulation + how multiplicity appears in the reformulation

Chat with @laradicp :

• modify and delete not supported during opitimization because cannot propagate change to reformulation

[codecov]

Codecov Report

Merging #564 (c62a22d) into master (6178b56) will decrease coverage by 0.48%.
The diff coverage is 100.00%.

❗ Current head c62a22d differs from pull request most recent head b6a9ecb. Consider uploading reports for the commit b6a9ecb to get more accurate results

@@            Coverage Diff             @@
##           master     #564      +/-   ##
==========================================
- Coverage   86.87%   86.39%   -0.49%     
==========================================
  Files          47       47              
  Lines        4723     4718       -5     
==========================================
- Hits         4103     4076      -27     
- Misses        620      642      +22     

Impacted Files	Coverage Δ
src/Algorithm/basic/solveipform.jl	89.47% <ø> (ø)
src/Algorithm/colgen.jl	95.73% <ø> (ø)
src/Algorithm/conquer.jl	81.81% <ø> (ø)
src/Algorithm/treesearch.jl	94.18% <ø> (ø)
src/Algorithm/basic/solvelpform.jl	91.42% <100.00%> (-2.86%)	⬇️
src/MathProg/formulation.jl	77.57% <100.00%> (ø)
src/MOIwrapper.jl	81.12% <0.00%> (-4.78%)	⬇️
src/MathProg/MOIinterface.jl	90.87% <0.00%> (-0.83%)	⬇️

---

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 b4a66a6...b6a9ecb. Read the comment docs.

[guimarqu]

@laradicp it looks like there is a problem with the disaggregated solution interface, may you look what's going on ?

This is the bug from the documentation job log :

┌ Warning: failed to run `@example` block in src/man/decomposition.md:152-166
│ ```@example identical
│ for t in T
│     assignment_patterns = BD.getsolutions(model, t);
│     for pattern in assignment_patterns
│         nb_times_pattern_used = BD.value(pattern);
│         jobs_in_pattern = [];
│         for j in J
│             if BD.value(pattern, x[t, j]) ≈ 1
│                 push!(jobs_in_pattern, j);
│             end
│         end
│         println("Pattern of machine type $t used $nb_times_pattern_used times : $jobs_in_pattern");
│     end
│ end
│ ```
│   c.value = BoundsError: attempt to access 1-element Vector{Coluna.ColunaBase.Solution{Coluna.MathProg.Formulation{Coluna.MathProg.DwMaster}, Coluna.MathProg.VarId, Float64}} at index [2]
└ @ Documenter.Expanders ~/.julia/packages/Documenter/R2HVS/src/Expanders.jl:591

[laradicp]

Yes, there's a problem with the usage of k, it's not referring correctly to the subproblem. I'm working on it and I think the fix is just in getsolutions. It worked with the test I made because the axis was unitary. Should I update the test too even if it takes more time to solve?

[guimarqu]

No I think I can add a doctest