Merge pull request #9 from MillironX/feature/variations-parse

Manifest.toml

@@ -23,9 +23,9 @@ version = "2.0.0"
 
 [[BioGenerics]]
 deps = ["TranscodingStreams"]
-git-tree-sha1 = "57deb413ca9f4c8bc7d4c6e98ebe217ff728c737"
+git-tree-sha1 = "6d3f3b474b3df2e83dc67ad12ec63aee4eb5241b"
 uuid = "47718e42-2ac5-11e9-14af-e5595289c2ea"
-version = "0.1.0"
+version = "0.1.1"
 
 [[BioSequences]]
 deps = ["BioGenerics", "BioSymbols", "Combinatorics", "IndexableBitVectors", "Printf", "Random", "StableRNGs", "Twiddle"]
@@ -50,6 +50,10 @@ git-tree-sha1 = "8cd7b7d1c7f6fcbe7e8743a58adf57788ec7f787"
 uuid = "34da2185-b29b-5c13-b0c7-acf172513d20"
 version = "3.18.0"
 
+[[CompilerSupportLibraries_jll]]
+deps = ["Artifacts", "Libdl"]
+uuid = "e66e0078-7015-5450-92f7-15fbd957f2ae"
+
 [[DataStructures]]
 deps = ["Compat", "InteractiveUtils", "OrderedCollections"]
 git-tree-sha1 = "0347f23484a96d56e7096eb1f55c6975be34b11a"
@@ -69,9 +73,12 @@ deps = ["Random", "Serialization", "Sockets"]
 uuid = "8ba89e20-285c-5b6f-9357-94700520ee1b"
 
 [[Downloads]]
-deps = ["ArgTools", "LibCURL", "NetworkOptions"]
+deps = ["ArgTools", "FileWatching", "LibCURL", "NetworkOptions"]
 uuid = "f43a241f-c20a-4ad4-852c-f6b1247861c6"
 
+[[FileWatching]]
+uuid = "7b1f6079-737a-58dc-b8bc-7a2ca5c1b5ee"
+
 [[IndexableBitVectors]]
 deps = ["Random", "Test"]
 git-tree-sha1 = "b7f5e42dc867b8a8654a5f899064632dac05bc82"
@@ -108,7 +115,7 @@ uuid = "29816b5a-b9ab-546f-933c-edad1886dfa8"
 uuid = "8f399da3-3557-5675-b5ff-fb832c97cbdb"
 
 [[LinearAlgebra]]
-deps = ["Libdl"]
+deps = ["Libdl", "libblastrampoline_jll"]
 uuid = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 
 [[Logging]]
@@ -131,13 +138,17 @@ uuid = "14a3606d-f60d-562e-9121-12d972cd8159"
 [[NetworkOptions]]
 uuid = "ca575930-c2e3-43a9-ace4-1e988b2c1908"
 
+[[OpenBLAS_jll]]
+deps = ["Artifacts", "CompilerSupportLibraries_jll", "Libdl"]
+uuid = "4536629a-c528-5b80-bd46-f80d51c5b363"
+
 [[OrderedCollections]]
 git-tree-sha1 = "16c08bf5dba06609fe45e30860092d6fa41fde7b"
 uuid = "bac558e1-5e72-5ebc-8fee-abe8a469f55d"
 version = "1.3.1"
 
 [[Pkg]]
-deps = ["Artifacts", "Dates", "Downloads", "LibGit2", "Libdl", "Logging", "Markdown", "Printf", "REPL", "Random", "SHA", "Serialization", "TOML", "Tar", "UUIDs"]
+deps = ["Artifacts", "Dates", "Downloads", "LibGit2", "Libdl", "Logging", "Markdown", "Printf", "REPL", "Random", "SHA", "Serialization", "TOML", "Tar", "UUIDs", "p7zip_jll"]
 uuid = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
 
 [[Printf]]
@@ -153,7 +164,7 @@ deps = ["InteractiveUtils", "Markdown", "Sockets", "Unicode"]
 uuid = "3fa0cd96-eef1-5676-8a61-b3b8758bbffb"
 
 [[Random]]
-deps = ["Serialization"]
+deps = ["SHA", "Serialization"]
 uuid = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 
 [[SHA]]
@@ -217,6 +228,14 @@ uuid = "4ec0a83e-493e-50e2-b9ac-8f72acf5a8f5"
 deps = ["Libdl"]
 uuid = "83775a58-1f1d-513f-b197-d71354ab007a"
 
+[[libblastrampoline_jll]]
+deps = ["Artifacts", "Libdl", "OpenBLAS_jll"]
+uuid = "8e850b90-86db-534c-a0d3-1478176c7d93"
+
 [[nghttp2_jll]]
 deps = ["Artifacts", "Libdl"]
 uuid = "8e850ede-7688-5339-a07c-302acd2aaf8d"
+
+[[p7zip_jll]]
+deps = ["Artifacts", "Libdl"]
+uuid = "3f19e933-33d8-53b3-aaab-bd5110c3b7a0"

Project.toml

@@ -5,5 +5,14 @@ version = "0.1.0"
 
 [deps]
 BioAlignments = "00701ae9-d1dc-5365-b64a-a3a3ebf5695e"
+BioGenerics = "47718e42-2ac5-11e9-14af-e5595289c2ea"
 BioSequences = "7e6ae17a-c86d-528c-b3b9-7f778a29fe59"
 BioSymbols = "3c28c6f8-a34d-59c4-9654-267d177fcfa9"
+
+[extras]
+BioAlignments = "00701ae9-d1dc-5365-b64a-a3a3ebf5695e"
+BioSequences = "7e6ae17a-c86d-528c-b3b9-7f778a29fe59"
+Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
+
+[targets]
+test = ["Test", "BioSequences", "BioAlignments"]

src/SequenceVariation.jl

@@ -20,9 +20,10 @@ TODO now:
 * Add tests
 """
 
-using BioSymbols: BioSymbol
 using BioAlignments: BioAlignments, PairwiseAlignment
+using BioGenerics: BioGenerics, leftposition
 using BioSequences: BioSequences, BioSequence, NucleotideSeq, AminoAcidSeq, LongSequence, isgap
+using BioSymbols: BioSymbol
 
 const BA = BioAlignments
 const BS = BioSequences
@@ -133,6 +134,9 @@ function Base.parse(::Type{<:Edit{Se, Sy}}, s::Union{String, SubString{String}})
     end
 end
 
+mutation(e::Edit) = e.x
+BioGenerics.leftposition(e::Edit) = e.pos
+
 #=
 @noinline throw_parse_error(T, p::Integer) = error("Failed to parse $T at byte $p")
 
@@ -341,6 +345,18 @@ end
 
 Variation(ref::S, edit::Edit{S, T}) where {S, T} = Variation{S, T}(ref, edit)
 
+function Variation(ref::S, edit::AbstractString) where {S<:BioSequence}
+    T = eltype(ref)
+
+    e = parse(Edit{S,T}, edit)
+    return Variation{S,T}(ref, e)
+end
+
+reference(v::Variation) = v.reference
+edit(v::Variation) = v.edit
+mutation(v::Variation) = mutation(edit(v))
+BioGenerics.leftposition(v::Variation) = leftposition(edit(v))
+
 function is_valid(v::Variation)
     isempty(v.ref) && return false
     op = v.edit.x
@@ -426,6 +442,8 @@ export Insertion,
     Deletion,
     Substitution,
     Variant,
-    Variation
+    Variation,
+    reference,
+    mutation
 
 end # module

test/runtests.jl

@@ -23,9 +23,10 @@ TODO now:
 * Add tests
 """
 
-using BioSequences
 using BioAlignments
+using BioSequences
 using SequenceVariation
+using Test
 
 const DNA_MODEL = BioAlignments.AffineGapScoreModel(EDNAFULL, gap_open=-25, gap_extend=-2)
 
@@ -44,3 +45,15 @@ var = Variant(align(seq1, seq2))
     aln02 = PairwiseAlignment(read02, refseq)
     @test Variant(aln01).edits == Variant(aln02).edits
 end
+
+@testset "VariationParsing" begin
+    refseq = dna"ACAACTTTATCT"
+
+    sub = Variation(refseq, "A4T")
+    del = Variation(refseq, "Δ4-5")
+    ins = Variation(refseq, "4TT")
+
+    @test mutation(sub) isa Substitution
+    @test mutation(del) isa Deletion
+    @test mutation(ins) isa Insertion
+end