Allow join to work with BioSymbol (#234)

Previously, it was only possible to join BioSequences. Now, one can call
join(LongRNA{2}, (RNA_U, rna"AGCA")) and have it work.
Also skip unneeded calls to resize! when joining

Author: jakobnissen

CHANGELOG.md

@@ -4,6 +4,11 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
 and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
+## [UNRELEASED]
+
+### Added
+* It is now possible to `join` BioSymbols into a BioSequence.
+
 ## [3.0.1]
 ### Removed
 

src/biosequence/biosequence.jl

@@ -100,13 +100,13 @@ end
 
 # Fast path for iterables we know are stateless
 function join!(seq::BioSequence, it::Union{Vector, Tuple, Set})
-    _join!(resize!(seq, sum(length, it, init=0)), it, Val(true))
+    _join!(resize!(seq, sum(joinlen, it, init=0)), it, Val(true))
 end
 
 """
     join!(seq::BioSequence, iter)
 
-Concatenate all biosequences in `iter` into `seq`, resizing it to fit.
+Concatenate all biosequences/biosymbols in `iter` into `seq`, resizing it to fit.
 
 # Examples
 ```
@@ -122,19 +122,27 @@ join!(seq::BioSequence, it) = _join!(seq, it, Val(false))
 # B is whether the size of the destination seq is already
 # known to be the final size
 function _join!(seq::BioSequence, it, ::Val{B}) where B
-    index = 1
+    len = 0
+    oldlen = length(seq)
     for i in it
-        B || resize!(seq, length(seq) + length(i))
-        copyto!(seq, index, i, 1, length(i))
-        index += length(i)
+        pluslen = joinlen(i)
+        if !B && oldlen < (len + pluslen)
+            resize!(seq, len + pluslen)
+        end
+        if i isa BioSymbol
+            seq[len + 1] = i
+        else
+            copyto!(seq, len + 1, i, 1, length(i))
+        end
+        len += pluslen
     end
     seq
 end
 
 """
     join(::Type{T <: BioSequence}, seqs)
 
-Concatenate all the `seqs` to a biosequence of type `T`.
+Concatenate all the biosequences/biosymbols in `seqs` to a biosequence of type `T`.
 
 # Examples
 ```
@@ -146,9 +154,12 @@ TAGAAC
 see also [`join!`](@ref)
 """
 function Base.join(::Type{T}, it::Union{Vector, Tuple, Set}) where {T <: BioSequence}
-    _join!(T(undef, sum(length, it, init=0)), it, Val(true))
+    _join!(T(undef, sum(joinlen, it, init=0)), it, Val(true))
 end
 
+# length is intentionally not implemented for BioSymbol
+joinlen(x::Union{BioSequence, BioSymbol}) = x isa BioSymbol ? 1 : length(x)
+
 function Base.join(::Type{T}, it) where {T <: BioSequence}
     _join!(empty(T), it, Val(false))
 end

test/biosequences/biosequence.jl

@@ -49,10 +49,16 @@ random_simple(len::Integer) = SimpleSeq(rand([RNA_A, RNA_C, RNA_G, RNA_U], len))
 
     seq2 = SimpleSeq([RNA_U, RNA_C, RNA_U])
     gen = (i for i in [seq, seq2])
-    @test join!(SimpleSeq([]), [seq, seq2]) == SimpleSeq([RNA(i) for i in "CGUUCU"])
-    @test join!(SimpleSeq([]), gen) == SimpleSeq([RNA(i) for i in "CGUUCU"])
+    newseq = SimpleSeq([])
+    join!(newseq, [seq, seq2])
+    @test newseq == SimpleSeq([RNA(i) for i in "CGUUCU"]) 
+    join!(newseq, gen)
+    @test newseq == SimpleSeq([RNA(i) for i in "CGUUCU"])
+    join!(newseq, [RNA_U, RNA_C, SimpleSeq([RNA_G, RNA_C])])
+    @test newseq == SimpleSeq([RNA(i) for i in "UCGC"])
     @test join(SimpleSeq, [seq, seq2]) == join!(SimpleSeq([]), [seq, seq2])
     @test join(SimpleSeq, gen) == join!(SimpleSeq([]), gen)
+    @test join(SimpleSeq, [RNA_U, RNA_G, seq, RNA_U]) == SimpleSeq([RNA(i) for i in "UGCGUU"]) 
 
     @test copy!(SimpleSeq([]), seq) == seq
     seq3 = copy(seq2)

test/longsequences/basics.jl

@@ -263,6 +263,11 @@ end
 
     base_aa_seq = aa"KMAEEHPAIYWLMN"
     test_join(LongAA, (aa"KMVLE", aa"", (@view base_aa_seq[3:6])), aa"KMVLEAEEH")
+
+    # Joining seqs and symbols
+    test_join(LongAA, [AA_G, AA_P, AA_L, aa"MNVWEED", AA_K], aa"GPLMNVWEEDK")
+    test_join(LongRNA{4}, [RNA_M, RNA_U, RNA_S, rna"AGCGSK"], rna"MUSAGCGSK")
+    test_join(LongDNA{2}, [dna"ATGCTTA", DNA_G, DNA_G, DNA_A, DNA_A, DNA_A], dna"ATGCTTAGGAAA")
 end
 
 @testset "Length" begin