This would make it less frustrating for users to specify input to functions

Hello,

Thank you for the great code!
Through working with it, I have noticed two points to be fixed in relation to randomized ordered list generation.

randomize.jl line 30
sort!(cpd_masses, by= x->x[2], rev=true)
This will sort compounds in descending order of molecular mass.

randomize.jl line 56

i,j = rand(1:length(tuples),2) ## 2 random indices 
if i > j
	i,j = j,i
end

By adding the constraint, the tendency for compounds to be ordered by molecular mass is maintained after 10,000 exchange operations. Although Handorf et al. 2008 did not describe the algorithm for this part in detail, it seems that they added the constraint.

Thank you!
Hayate

1. DomainError incorrectly thrown
   In function
   function find_minimal_seed_set( rstructs::Reactions, rids::IDs, sids::IDs; tids::IDs=IDs(), write_path::Union{String,Nothing}=nothing, allowed_forward::Union{SparseVector{Bool,Int64},Nothing}=nothing, allowed_backward::Union{SparseVector{Bool,Int64},Nothing}=nothing)

the Vector(x) != ones(Int,length(cids)) && throw(DomainError("This should be a vector of all 1s of length(cids")) error was thrown, even though conditions were satisfied. I commented this out in my copy.

2. Nothing as a subscript
   In function function loop_and_remove_seeds( sids::IDs, cids::IDs, x::SparseVector{Int64,Int64}, t::SparseVector{Int64,Int64}, R::SparseMatrixCSC{Int64,Int64}, P::SparseMatrixCSC{Int64,Int64}; allowed_forward::Union{SparseVector{Bool,Int64},Nothing}=nothing, allowed_backward::Union{SparseVector{Bool,Int64},Nothing}=nothing, skip_seed_indices::Int=0)

the for i in seed_indicies loop, specifically the x[i] = 0 line, tries to use nothing as i. To fix this, I added a line within the for loop, if i != nothing, and included all elements of the loop inside this if block.

3. Simple write out file path
   In function function simple_write_out( path::String, x::Union{Vector{Int},SparseVector{Int64,Int64}}, t::Union{Vector{Int},SparseVector{Int64,Int64}}, cids::IDs, rids::IDs, X::Union{Vector{Vector{Int}},Vector{SparseVector{Int64,Int64}}}, Y::Union{Vector{Vector{Int}},Vector{SparseVector{Int64,Int64}}})

there is no designated file path - only a directory. I added a json extension to the open command. open(path*"output.json","w") do f

to make sure the loop successfully removes compounds when using our sparse construction

Need to add a script to create following mappings under links folder,

• pathway_bioinformation ==> map or rn number : [class, subclass, name]

• formula_compound ==> chemical formula : [C00001, C00002, ...]

• atom_compound ==> atom : [C00001, C00004, ...]