Move tutorial to site

docs/src/tutorial.md

@@ -1,114 +1,3 @@
 # Tutorial
 
-## Input
-
-RipQP uses the package [QuadraticModels.jl](https://github.com/JuliaSmoothOptimizers/QuadraticModels.jl) to model
-convex quadratic problems.
-
-Here is a basic example:
-
-```@example QM
-using QuadraticModels, LinearAlgebra, SparseMatricesCOO
-Q = [6. 2. 1.
-    2. 5. 2.
-    1. 2. 4.]
-c = [-8.; -3; -3]
-A = [1. 0. 1.
-    0. 2. 1.]
-b = [0.; 3]
-l = [0.;0;0]
-u = [Inf; Inf; Inf]
-QM = QuadraticModel(c, SparseMatrixCOO(tril(Q)), A=SparseMatrixCOO(A), lcon=b, ucon=b, 
-                    lvar=l, uvar=u, c0=0., name="QM")
-```
-
-Once your `QuadraticModel` is loaded, you can simply solve it RipQP:
-
-```@example QM
-using RipQP
-stats = ripqp(QM)
-println(stats)
-```
-
-The `stats` output is a
-[GenericExecutionStats](https://juliasmoothoptimizers.github.io/SolverCore.jl/dev/reference/#SolverCore.GenericExecutionStats).
-
-It is also possible to use the package [QPSReader.jl](https://github.com/JuliaSmoothOptimizers/QPSReader.jl) in order to
-read convex quadratic problems in MPS or SIF formats:
-
-```julia
-using QPSReader, QuadraticModels
-QM = QuadraticModel(readqps("QAFIRO.SIF"))
-```
-
-## Logging
-
-RipQP displays some logs at each iterate.
-
-You can deactivate logging with
-
-```julia
-stats = ripqp(QM, display = false)
-```
-
-It is also possible to get a history of several quantities such as the primal and dual residuals and the relative primal-dual gap.
-These quantites are available in the dictionnary `solver_specific` of the `stats`.
-
-```julia
-stats = ripqp(QM, history = true)
-pddH = stats.solver_specific[:pddH]
-```
-
-## Change configuration and tolerances
-
-You can use `RipQP` without scaling with:
-
-```julia
-stats = ripqp(QM, scaling = false)
-```
-
-You can also change the [`RipQP.InputTol`](@ref) type to change the tolerances for the stopping criteria:
-
-```julia
-stats = ripqp(QM, itol = InputTol(max_iter = 100, ϵ_rb = 1.0e-4), scaling = false)
-```
-
-## Save the Interior-Point system
-
-At every iteration, RipQP solves two linear systems with the default Predictor-Corrector method (the affine system and the corrector-centering system), or one linear system with the Infeasible Path-Following method.
-
-To save these systems, you can use:
-
-```julia
-w = SystemWrite(write = true, name="test_", kfirst = 4, kgap=3) 
-stats1 = ripqp(QM, w = w)
-```
-
-This will save one matrix and the associated two right-hand sides of the PC method every three iterations starting at iteration four.
-Then, you can read the saved files with:
-
-```julia
-using DelimitedFiles, MatrixMarket
-K = MatrixMarket.mmread("test_K_iter4.mtx")
-rhs_aff = readdlm("test_rhs_iter4_aff.rhs", Float64)[:]
-rhs_cc = readdlm("test_rhs_iter4_cc.rhs", Float64)[:] 
-```
-
-## Timers
-
-You can see the elapsed time with:
-
-```julia
-stats1.elapsed_time
-```
-
-For more advanced timers you can use [`TimerOutputs.jl`](https://github.com/KristofferC/TimerOutputs.jl):
-
-```julia
-using TimerOutputs
-TimerOutputs.enable_debug_timings(RipQP)
-reset_timer!(RipQP.to)
-stats = ripqp(QM)
-TimerOutputs.complement!(RipQP.to) # print complement of timed sections
-show(RipQP.to, sortby = :firstexec)
-```
+Check an [introduction to RipQP](https://jso-docs.github.io/introduction-to-ripqp/) and more tutorials on the [JSO tutorials page](https://juliasmoothoptimizers.github.io/tutorials/).