Docs: `approximate` page shows a keyword-argument signature that doesn't seem to exist

[VinceNeede]

While looking at the approximate documentation, I noticed it suggests the function can be called with keyword arguments, but I couldn't find a matching method — happy to be told I'm missing something!

julia> methods(approximate)
# 8 methods for generic function "approximate" from MPSKit:
 [1] approximate(mps::MultilineMPS, toapprox::Tuple{MultilineMPO, MultilineMPS}, alg::VOMPS, envs)
     @ MPSKit src/algorithms/approximate/vomps.jl:13
 [2] approximate(ψ::MultilineMPS, toapprox::Tuple{MultilineMPO, MultilineMPS}, alg::VUMPS, envs...; kwargs...)
     @ deprecated.jl:213
 [3] approximate(mps::MultilineMPS, toapprox::Tuple{MultilineMPO, MultilineMPS}, alg::VOMPS)
     @ MPSKit src/algorithms/approximate/vomps.jl:13
 [4] approximate(ψ::InfiniteMPS, toapprox::Tuple{InfiniteMPO, InfiniteMPS}, algorithm::Union{IDMRG, IDMRG2})
     @ MPSKit src/algorithms/approximate/approximate.jl:61
 [5] approximate(ψ::InfiniteMPS, toapprox::Tuple{InfiniteMPO, InfiniteMPS}, algorithm)
     @ MPSKit src/algorithms/approximate/approximate.jl:39
 [6] approximate(ψ::InfiniteMPS, toapprox::Tuple{InfiniteMPO, InfiniteMPS}, algorithm::Union{IDMRG, IDMRG2}, envs)
     @ MPSKit src/algorithms/approximate/approximate.jl:61
 [7] approximate(ψ::InfiniteMPS, toapprox::Tuple{InfiniteMPO, InfiniteMPS}, algorithm, envs)
     @ MPSKit src/algorithms/approximate/approximate.jl:39
 [8] approximate(ψ, toapprox, alg::Union{DMRG, DMRG2, IDMRG, IDMRG2}, envs...)
     @ MPSKit src/algorithms/approximate/approximate.jl:54

The only method that accepts kwargs... [2] is marked @ deprecated.jl, so it doesn't look like the intended public signature either. It might just be that the docs are slightly out of sync with the current method signatures above — or maybe keyword-argument support was planned/removed at some point and the docs weren't updated to match.

I'd be happy to take a look and open a PR — just let me know which direction is preferred: dropping the explicit algorithm argument in favor of keyword-based dispatch (matching what the docs currently describe), or updating the docs to align with the existing method signatures.

[lkdvos]

Thanks for reporting this!

In general my preferred solution would be something similar to find_groundstate:

    find_groundstate(ψ₀, H, [environments]; kwargs...) -> (ψ, environments, ϵ)
    find_groundstate(ψ₀, H, algorithm, environments) -> (ψ, environments, ϵ)

where the first signature attempts to come up with a sensible default algorithm based on the supplied keywords, and the second performs the actual implementations.
The hard part is usually figuring out the sensible default algorithm though 🙃

If you want to have a look and suggest something, very welcome!