Add basic presolve into MatOI

I open this issue to discuss a more long-term subject: I wonder if we could include basic presolve operations into MatOI?
I think it would be easier to develop that inside this package instead of MOI, but I may be wrong. The idea would be to share the efforts and get good presolve abilities for LP's solver written in pure Julia, like [Tulip.jl](https://github.com/ds4dm/Tulip.jl/) or [Simplex.jl](https://github.com/exanauts/Simplex.jl).

An idea would be to adapt already existing codes, like
- [Scipy presolve](https://github.com/scipy/scipy/blob/v1.5.1/scipy/optimize/_linprog_util.py#L443)
- [Tulip.jl presolve](https://github.com/ds4dm/Tulip.jl/tree/master/src/Presolve), cc @mtanneau
- [HiGHs presolve](https://github.com/ERGO-Code/HiGHS/tree/master/src/presolve)

I think [Presolving in linear programming](https://link.springer.com/article/10.1007/BF01586000) by E. Andersen and K. Andersen, is also a good paper to grasp inspiration from.

Start by comparing SciPy's scipy/optimize/_linprog_util.py, Tulip.jl's src/Presolve, and HiGHS's src/presolve, alongside the cited presolve paper. The issue needs a concrete scope and an agreed definition of basic presolve before implementation can be considered done.

由索引模型根据 Issue 内容生成。