Js/nc groebner basis documentation (#1739)

docs/src/free_associative_algebra.md

@@ -229,6 +229,19 @@ generators of that ideal.
 Since such a Groebner basis is not necessarily finite, one can additionally pass a `reduction_bound`
 to the function, to only compute a partial Groebner basis.
 
+```@docs
+groebner_basis(g::Vector{FreeAssAlgElem{T}}, reduction_bound::Int = typemax(Int), remove_redundancies::Bool = false) where T <: FieldElement
+
+normal_form(f::FreeAssAlgElem{T}, g::Vector{FreeAssAlgElem{T}}, aut::AhoCorasickAutomaton) where T
+
+interreduce!(g::Vector{FreeAssAlgElem{T}}) where T
+```
+
+The implementation uses a non-commutative version of the Buchberger algorithm as described in
+> Xingqiang Xiu,
+> Non-commutative Gröbner Bases and Applications,
+> PhD thesis, 2012.
+
 **Examples**
 
 ```jldoctest; setup = :(using AbstractAlgebra)

src/generic/FreeAssAlgebraGroebner.jl

@@ -624,6 +624,9 @@ Compute a Groebner basis for the ideal generated by `g`. Stop when `reduction_bo
 non-zero entries have been added to the Groebner basis. If the computation stops due to the bound being exceeded, 
 the result is in general not an actual Groebner basis, just a subset of one. However, whenever the normal form with
 respect to this incomplete Groebner basis is `0`, it will also be `0` with respect to the full Groebner basis.
+
+If `remove_redundancies` is set to true, some redundant obstructions will be removed during the computation, which might save
+time, however in practice it seems to inflate the running time regularly.
 """ 
 function groebner_basis(
     g::Vector{FreeAssAlgElem{T}},