Port to Julia 1.11: Graphs.jl, modern CSV, AbstractString types

- Replace LightGraphs (archived) with Graphs.jl (drop-in replacement)
- Remove CSV 0.8.5 pin, allow CSV 0.10+
- Change String -> AbstractString in distribution signatures to handle
  CSV.jl's String3/String31 types (from PR probcomp#26 by limarta)
- Rename proposal_dummy_value -> PROPOSAL_DUMMY_VALUE (constant convention)
- Remove unused deps: BenchmarkTools, DelimitedFiles, Interpolations,
  Memoize, Plots, Polynomials, PyPlot, Revise, StatsBase
- Add [compat] bounds for all dependencies
- Tested: loads and precompiles on Julia 1.11.9

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: todonnell

Project.toml

@@ -4,23 +4,20 @@ authors = ["alex-lew <alexl.mail@gmail.com>"]
 version = "0.1.0"
 
 [deps]
-BenchmarkTools = "6e4b80f9-dd63-53aa-95a3-0cdb28fa8baf"
 CSV = "336ed68f-0bac-5ca0-87d4-7b16caf5d00b"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"
-DelimitedFiles = "8bb1440f-4735-579b-a4ab-409b98df4dab"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
-Interpolations = "a98d9a8b-a2ab-59e6-89dd-64a1c18fca59"
+Graphs = "86223c79-3864-5bf0-83f7-82e725a168b6"
 JSON = "682c06a0-de6a-54ab-a142-c8b1cf79cde6"
-LightGraphs = "093fc24a-ae57-5d10-9952-331d41423f4d"
 MacroTools = "1914dd2f-81c6-5fcd-8719-6d5c9610ff09"
-Memoize = "c03570c3-d221-55d1-a50c-7939bbd78826"
-Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
-Polynomials = "f27b6e38-b328-58d1-80ce-0feddd5e7a45"
-PyPlot = "d330b81b-6aea-500a-939a-2ce795aea3ee"
-Revise = "295af30f-e4ad-537b-8983-00126c2a3abe"
-StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
 StringDistances = "88034a9c-02f8-509d-84a9-84ec65e18404"
 
 [compat]
-CSV = "0.8.5"
+julia = "1.10"
+CSV = "0.10"
+DataFrames = "1"
+Distributions = "0.25"
+Graphs = "1"
+MacroTools = "0.5"
+StringDistances = "0.11"

src/PClean.jl

@@ -1,7 +1,7 @@
 module PClean
 
 using Distributions
-using LightGraphs
+using Graphs
 using CSV
 using DataFrames: DataFrame
 

src/distributions/add_typos.jl

@@ -33,7 +33,7 @@ end
 
 const IMPOSSIBLE = -1e5
 
-random(::AddTypos, word::String, max_typos=nothing) = begin
+random(::AddTypos, word::AbstractString, max_typos=nothing) = begin
   num_typos = rand(NegativeBinomial(ceil(length(word) / 5.0), 0.9))
   num_typos = isnothing(max_typos) ? num_typos : min(max_typos, num_typos)
   for i=1:num_typos
@@ -47,7 +47,7 @@ end
 const add_typos_density_dict = Dict{Tuple{String, String}, Float64}()
 const LETTERS_PER_TYPO = 5.0
 
-logdensity(::AddTypos, observed::Union{String,Missing}, word::String, max_typos=nothing) = begin
+logdensity(::AddTypos, observed::Union{<:AbstractString,Missing}, word::AbstractString, max_typos=nothing) = begin
   if ismissing(observed)
     return 0.0
   end

src/distributions/distributions.jl

@@ -5,7 +5,7 @@ function random end
 function logdensity end
 
 struct ProposalDummyValue end
-const proposal_dummy_value = ProposalDummyValue()
+const PROPOSAL_DUMMY_VALUE = ProposalDummyValue()
 
 # Can this distribution enumerate values on which
 # its posterior support is concentrated? If the distribution's

src/distributions/string_prior.jl

@@ -13,19 +13,19 @@ const alphabet_lookup = Dict([l => i for (i, l) in enumerate(alphabet)])
 has_discrete_proposal(::StringPrior) = true
 
 # Assume proposal_atoms are unique.
-function discrete_proposal(::StringPrior, min_length::Int, max_length::Int, proposal_atoms::Vector{String})::Tuple{Vector{Union{String, ProposalDummyValue}}, Vector{Float64}}
-  options = Union{String, ProposalDummyValue}[proposal_atoms..., proposal_dummy_value]
+function discrete_proposal(::StringPrior, min_length::Int, max_length::Int, proposal_atoms::Vector{T})::Tuple{Vector{Union{T, ProposalDummyValue}}, Vector{Float64}} where T<:AbstractString
+  options = Union{T, ProposalDummyValue}[proposal_atoms..., PROPOSAL_DUMMY_VALUE]
   probs = map(s -> logdensity(StringPrior(), s, min_length, max_length, proposal_atoms), proposal_atoms)
   total = logsumexp(probs)
   probs = Float64[probs..., log1p(-exp(total))]
   return (options, probs)
 end
 
-discrete_proposal_dummy_value(::StringPrior, min_length::Int, max_length::Int, proposal_atoms::Vector{String}) = begin
+discrete_proposal_dummy_value(::StringPrior, min_length::Int, max_length::Int, proposal_atoms::Vector{<:AbstractString}) = begin
   join(fill("*", Int(floor((min_length + max_length) / 2))))
 end
 
-random(::StringPrior, min_length::Int, max_length::Int, proposal_atoms::Vector{String}) = begin
+random(::StringPrior, min_length::Int, max_length::Int, proposal_atoms::Vector{<:AbstractString}) = begin
   len = rand(DiscreteUniform(min_length, max_length))
   letters = []
   for i=1:len
@@ -40,7 +40,7 @@ end
 
 const UNUSUAL_LETTER_PENALTY = 1000
 const string_prior_density_dict = Dict{Tuple{String, Int, Int}, Float64}()
-function logdensity(::StringPrior, observed::String, min_length::Int, max_length::Int, proposal_atoms::Vector{String})
+function logdensity(::StringPrior, observed::AbstractString, min_length::Int, max_length::Int, proposal_atoms::Vector{<:AbstractString})
   get!(string_prior_density_dict, (observed, min_length, max_length)) do
     if length(observed) < min_length || length(observed) > max_length
       return -Inf

src/distributions/time_prior.jl

@@ -6,7 +6,7 @@ has_discrete_proposal(::TimePrior) = true
 
 # Assume proposal_atoms are unique.
 function discrete_proposal(::TimePrior, proposal_atoms::Vector{String})
-  options = [proposal_atoms..., proposal_dummy_value]
+  options = [proposal_atoms..., PROPOSAL_DUMMY_VALUE]
   probs = map(x -> isnothing(match(r"^\d?\d:\d\d [ap]\.m\.$", x)) ? -Inf : -log(1440), proposal_atoms)
   total = logsumexp(probs)
   probs = [probs..., log1p(-exp(total))]

src/dsl/builder.jl

@@ -1,4 +1,4 @@
-using LightGraphs
+using Graphs
 
 
 ############

src/utils.jl

@@ -30,7 +30,7 @@ end
 
 Like `normalize`, but in log-space and more numerically stable.
 """
-logprobs(proportions::Vector{Float64}) = begin
+logprobs(proportions::AbstractVector{<:Real}) = begin
   l = log.(proportions)
   #l .- logsumexp(l)
 end