support multithreading

.gitignore

@@ -1,4 +1,6 @@
 /Manifest.toml
 /docs/build/
 /db
+/test/db
+/test/taxdump.tar.gz
 /.vscode

Project.toml

@@ -5,10 +5,13 @@ version = "0.4.2"
 
 [deps]
 AbstractTrees = "1520ce14-60c1-5f80-bbc7-55ef81b5835c"
+CodecZlib = "944b1d66-785c-5afd-91f1-9de20f533193"
 DataAPI = "9a962f9c-6df0-11e9-0e5d-c546b8b5ee8a"
+Downloads = "f43a241f-c20a-4ad4-852c-f6b1247861c6"
 InteractiveUtils = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
 OrderedCollections = "bac558e1-5e72-5ebc-8fee-abe8a469f55d"
 StringDistances = "88034a9c-02f8-509d-84a9-84ec65e18404"
+Tar = "a4e569a6-e804-4fa4-b0f3-eef7a1d5b13e"
 
 [compat]
 AbstractTrees = "0.4.3"

src/database.jl

@@ -36,21 +36,20 @@ function importnodes(nodes_dmp_path::String; db_size::Int=default_db_size)
     ranks = Vector{Symbol}(undef, db_size)
 
     c = 0
-    f = open(nodes_dmp_path, "r")
-    @inbounds(for line in eachline(f)
-        cols = split(line, "\t", limit=6)
-        cols[1] == cols[3] && continue
+    open(nodes_dmp_path, "r") do f
+        for line in eachline(f)
+            cols = split(line, "\t", limit=6)
 
-        taxid = parse(Int, cols[1])
-        parent = parse(Int, cols[3])
-        rank = Symbol(cols[5])
+            taxid = parse(Int, cols[1])
+            parent = parse(Int, cols[3])
+            rank = Symbol(cols[5])
 
-        c += 1
-        taxids[c] = taxid
-        parents[c] = parent
-        ranks[c] = rank
-    end)
-    close(f)
+            c += 1
+            @inbounds taxids[c] = taxid
+            @inbounds parents[c] = parent
+            @inbounds ranks[c] = rank
+        end
+    end
     resize!(taxids, c)
     resize!(parents, c)
     resize!(ranks, c)

src/taxon.jl

@@ -86,7 +86,7 @@ Return the `Taxon` object that is the parent of the given `Taxon` object.
 """
 function AbstractTrees.parent(taxon::Taxon)
    parent_taxid = get(taxon.db.parents, taxon.taxid, nothing)
-   if parent_taxid === nothing
+   if parent_taxid == taxon.taxid
         return nothing
    end
    parent = Taxon(parent_taxid, taxon.db)

test/runtests.jl

@@ -1,16 +1,25 @@
 using Taxonomy
 using Taxonomy.AbstractTrees
+using Downloads
+using Tar
+using CodecZlib
 using Test
 
+if !isdir("./db")
+    Downloads.download("ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/taxdump.tar.gz", "./taxdump.tar.gz")
+    tar_gz = open("./taxdump.tar.gz")
+    tar = GzipDecompressorStream(tar_gz)
+    Tar.extract(tar, "./db")
+    close(tar)
+else
+    @warn "Start test with existing database"
+end
+
 @testset "dabase.jl" begin
     @test isnothing(current_db())
 
     db = Taxonomy.DB("db/nodes.dmp", "db/names.dmp")
     @test current_db() == db
-
-    db_1 = Taxonomy.DB("db_1/nodes.dmp", "db_1/names.dmp")
-    current_db!(db_1)
-    @test current_db() == db_1
 end
 
 db = Taxonomy.DB("db/nodes.dmp", "db/names.dmp")