From 2f4b82cde5e67c20f23c694e6786f040d29e59cb Mon Sep 17 00:00:00 2001
From: Jakob Nybo Nissen <jakobnybonissen@gmail.com>
Date: Fri, 29 Dec 2023 11:06:15 +0100
Subject: [PATCH] Optimise complementation of nucleotides (#62)

While BioSequences complement nucleotides in bulk, making this necessity not
important, the upcoming Kmers.jl relies on efficient complementation of single
nucleotides, so this is important.
---
 src/nucleicacid.jl | 12 ++++++------
 test/runtests.jl   | 24 +++++++++++-------------
 2 files changed, 17 insertions(+), 19 deletions(-)

diff --git a/src/nucleicacid.jl b/src/nucleicacid.jl
index 1467ec4..7e2e423 100644
--- a/src/nucleicacid.jl
+++ b/src/nucleicacid.jl
@@ -427,12 +427,12 @@ RNA_A
 
 ```
 """
-function complement(nt::NucleicAcid)
-    bits = compatbits(nt)
-    return encode(
-        typeof(nt),
-        (bits & 0x01) << 3 | (bits & 0x08) >> 3 |
-        (bits & 0x02) << 1 | (bits & 0x04) >> 1)
+function complement(nt::Union{DNA, RNA})
+    # This is essentially a lookup table of 16 x 4 bits.
+    # It's the concatenation of the bitpatterns of the nucleotides,
+    # in order, complemented.
+    u64 = 0xf7b3d591e6a2c480 >>> ((4 * encoded_data(nt)) & 63)
+    reinterpret(typeof(nt), (u64 % UInt8) & 0x0f)
 end
 
 function Base.isvalid(::Type{T}, x::Integer) where T <: NucleicAcid
diff --git a/test/runtests.jl b/test/runtests.jl
index a80ef5d..706ab9c 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -239,19 +239,17 @@ end
     end
 
     @testset "complement" begin
-        @test complement(DNA_A) === DNA_T
-        @test complement(DNA_C) === DNA_G
-        @test complement(DNA_G) === DNA_C
-        @test complement(DNA_T) === DNA_A
-        @test complement(DNA_Gap) === DNA_Gap
-        @test complement(DNA_N) === DNA_N
-
-        @test complement(RNA_A) === RNA_U
-        @test complement(RNA_C) === RNA_G
-        @test complement(RNA_G) === RNA_C
-        @test complement(RNA_U) === RNA_A
-        @test complement(RNA_Gap) === RNA_Gap
-        @test complement(RNA_N) === RNA_N
+        for (a, b) in zip(
+            "-ACMGRSVTWYHKDBN",
+            "-TGKCYSBAWRDMHVN"
+        )
+            da, db = DNA(a), DNA(b)
+            for (i, j) in ((da, db), (RNA(da), RNA(db)))
+
+                @test complement(i) === j
+                @test complement(j) === i
+            end
+        end
     end
 
     @testset "Logic operations and Order" begin