Added support for the join operator, and more documentation

Project.toml

@@ -9,8 +9,8 @@ DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 GZip = "92fee26a-97fe-5a0c-ad85-20a5f3185b63"
 
 [compat]
-DataFrames = "0.19"
 BioSequences = "2"
+DataFrames = "0.19"
 julia = "1"
 
 [extras]

ci_prep.jl

@@ -0,0 +1,3 @@
+using Pkg.Registry
+Registry.add(Registry.RegistrySpec(url = "https://github.com/JuliaRegistries/General.git"))
+Registry.add(Registry.RegistrySpec(url = "https://github.com/BioJulia/BioJuliaRegistry.git"))

src/macro.jl

@@ -34,7 +34,9 @@ end
 Iterate over and evaluate expressions in `exs` for all genes in `chr.genes`,
 returning genes where all expressions evaluate to `true`. Any given symbol `s`
 in the expression will be substituted for `gene.s`. The gene itself can be
-accessed in the expression as `gene` (see example below).
+accessed in the expression as `gene`.
+
+# Examples
 ```julia
 julia> chromosome = readgbk("example.gbk")
 Chromosome 'example' (5028 bp) with 6 annotations

src/readgbk.jl

@@ -1,3 +1,6 @@
+"""
+Return the LOCUS entry of the header.
+"""
 function parseheader(header::String)
     lines = split(header, "\n")
     locus = split(lines[1], r" +")[2]
@@ -21,13 +24,14 @@ function parseposition(line::String)
     complete_left = !occursin('<', posstring)
     complete_right = !occursin('>', posstring)
     order = Vector{UnitRange{Int}}()
-    if occursin("order", posstring)
+	join = occursin("join", posstring)
+    if join || occursin("order", posstring)
         for m in eachmatch(r"\d+(\.\.|\^)\d+", posstring)
             r = Meta.parse.(split(m.match, r"(\.\.|\^)"))
             push!(order, r[1]:r[2])
         end
     end
-    return feature, Locus(position, strand, complete_left, complete_right, order)
+    return feature, Locus(position, strand, complete_left, complete_right, order, join)
 end
 
 

src/types.jl

@@ -3,14 +3,16 @@ abstract type AbstractGene end
 
 """
 Struct for storing information on genomic locations. `strand` can be '+', '-',
-or '.' when the strand is irrelevant.
+or '.' when the strand is irrelevant. `order` is used to store discontiguous
+sequences, indicated in the GenBank file with the order() and join() operators.
 """
 struct Locus
     position::UnitRange{Int}
     strand::Char
     complete_left::Bool
     complete_right::Bool
     order::Vector{UnitRange{Int}}
+    join::Bool
 end
 
 
@@ -20,10 +22,10 @@ end
     Locus(position::UnitRange{Int}, strand::Char)
 
 """
-Locus() = Locus(1:1, '.', true, true, UnitRange{Int}[])
-Locus(position::UnitRange{Int}) = Locus(position, '.', true, true, UnitRange{Int}[])
-Locus(position::UnitRange{Int}, strand::Char) = Locus(position, strand, true, true, UnitRange{Int}[])
-Locus(position::UnitRange{Int}, strand::Char, complete_left, complete_right) = Locus(position, strand, complete_left, complete_right, UnitRange{Int}[])
+Locus() = Locus(1:1, '.', true, true, UnitRange{Int}[], false)
+Locus(position::UnitRange{Int}) = Locus(position, '.', true, true, UnitRange{Int}[], false)
+Locus(position::UnitRange{Int}, strand::Char) = Locus(position, strand, true, true, UnitRange{Int}[], false)
+Locus(position::UnitRange{Int}, strand::Char, complete_left, complete_right) = Locus(position, strand, complete_left, complete_right, UnitRange{Int}[], false)
 
 Base.convert(::Type{Locus}, x::UnitRange{Int}) = Locus(x)
 function Base.convert(::Type{Locus}, x::StepRange{Int, Int})
@@ -35,6 +37,13 @@ function Base.convert(::Type{Locus}, x::StepRange{Int, Int})
     throw(DomainError(x, "`x` must have a step of 1 or -1"))
 end
 
+
+"""
+Struct for storing annotations for a single chromosome, plasmid, contig, etc.
+Contains five fields: `name`, `sequence`, `header`, `genes`, and `genedata`.
+Annotations are stored as a `DataFrame` in `genedata`, but can be accessed
+more easily through `genes` using the API provided in this module.
+"""
 mutable struct Chromosome{G <: AbstractGene}
     name::String
     sequence::LongDNASeq
@@ -59,7 +68,8 @@ end
 """
     addgene!(chr::Chromosome, feature, locus; kw...)
 
-Add gene to `chr`. `locus` can be a `Locus`, a UnitRange, or a StepRange.
+Add gene to `chr`. `locus` can be a `Locus`, a UnitRange, or a StepRange (for
+decreasing ranges, which will be annotated on the complementary strand).
 """
 function addgene!(chr::Chromosome, feature, locus; kw...)
     locus = convert(Locus, locus)
@@ -291,7 +301,11 @@ function Base.show(io::IO, locus::Locus)
     locus.strand == '-'     && (s *= "complement(")
     !locus.complete_left    && (s *= ">")
     if length(locus.order) > 0
-        s *= "order(" * join([join((r.start, r.stop), "..") for r in locus.order], ",") * ")"
+        if locus.join
+            s *= "join(" * join([join((r.start, r.stop), "..") for r in locus.order], ",") * ")"
+        else
+            s *= "order(" * join([join((r.start, r.stop), "..") for r in locus.order], ",") * ")"
+        end
     else
         s *= join((locus.position.start, locus.position.stop), "..")
     end
@@ -328,6 +342,11 @@ function formatsequence(sequence, io = IOBuffer)
 end
 
 
+"""
+    printgbk([io], chr)
+
+Print `chr` in GenBank format.
+"""
 function printgbk(chrs::AbstractVector{C}) where {C <: Chromosome}
     io = IOBuffer()
     printgbk(io, chrs)

test/runtests.jl

@@ -7,7 +7,7 @@ using Test
         s = "     gene            1..1"
         @test GenomicAnnotations.parseposition(s) == ("gene", Locus(1:1, '+'))
         s = "     gene            complement(order(3300..4037,4047..4052))"
-        @test GenomicAnnotations.parseposition(s) == ("gene", Locus(3300:4052, '-', true, true, UnitRange{Int}[3300:4037, 4047:4052]))
+        @test GenomicAnnotations.parseposition(s) == ("gene", Locus(3300:4052, '-', true, true, UnitRange{Int}[3300:4037, 4047:4052], false))
 
 
         chrs = readgbk("example.gbk")
@@ -71,7 +71,7 @@ using Test
     end
 
     @testset "Locus" begin
-        locus = Locus(1:1, '.', true, true, UnitRange{Int}[])
+        locus = Locus(1:1, '.', true, true, UnitRange{Int}[], false)
         @test Locus() == locus
         @test chr.genes[2].locus < chr.genes[4].locus
         @test chr.genes[2].locus == chr.genes[2].locus