Improve error when mis-encoding LongDNA from byte-like inputs (BioJul…

…ia#267)

This only applies to LongDNA encoded from byte-like inputs, since this uses a
fast path in BioSequences, where we have access to the whole sequence.
Other paths leading to the error happens from direct calls to e.g.
convert(DNA, x) in BioSymbols, so the error cannot be improved much.
Old error looks like: "Cannot encode 0x20 to DNAAlphabet{4}()".
New error looks like: "Cannot encode byte 0x20 (char ' ') at index 3 to DNAAlphabet{4}()".

CHANGELOG.md

@@ -5,6 +5,8 @@ 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
+* Improved error message when encoding LongDNA from byte-like objects
 
 ## [3.1.0]
 ### Added

src/longsequences/copying.jl

@@ -152,9 +152,19 @@ end
 # This is used to effectively scan an array of UInt8 for invalid bytes, when one is detected
 @noinline function throw_encode_error(A::Alphabet, src::AbstractArray{UInt8}, soff::Integer)
     for i in 1:div(64, bits_per_symbol(A))
-        sym = src[soff+i-1]
-        ascii_encode(A, sym) & 0x80 == 0x80 && error("Cannot encode $(repr(sym)) to $A")
+        index = soff + i - 1
+        sym = src[index]
+        if ascii_encode(A, sym) & 0x80 == 0x80
+            # If byte is a printable char, also display it
+            repr_char = if sym in UInt8('\a'):UInt8('\r') || sym in UInt8(' '):UInt8('~')
+                " (char '$(Char(sym))')"
+            else
+                ""
+            end
+            error("Cannot encode byte $(repr(sym))$(repr_char) at index $(index) to $A")
+        end
     end
+    @assert false "Expected error in encoding"
 end
 
 @inline function encode_chunk(A::Alphabet, src::AbstractArray{UInt8}, soff::Integer, N::Integer)

test/longsequences/basics.jl

@@ -24,6 +24,15 @@
     @test LongSequence(SimpleSeq("AUCGU")) isa LongRNA{2}
     @test LongSequence(SimpleSeq("AUCGU")) == LongRNA{2}("AUCGU")
     LongDNA{4}(LongRNA{4}("AUCGUA")) == LongDNA{4}("ATCGTA")
+
+    # Displays a nice error when constructed from strings substrings
+    # and bytearrays on encoding error
+    @static if VERSION >= v"1.8"
+        malformed = "ACWpNS"
+        @test_throws "Cannot encode byte $(repr(UInt8('p'))) (char 'p') at index 4 to BioSequences.DNAAlphabet{4}" LongDNA{4}(malformed)
+        malformed = "AGCUGUAGUCGGUAUAUAGGCGCGCUCGAUGAUGAUGCGUGCUGCUATDNANCUG"
+        @test_throws "Cannot encode byte $(repr(UInt8('T'))) (char 'T') at index $(length(malformed) - 7) to BioSequences.RNAAlphabet{2}" LongRNA{2}(malformed)
+    end
 end
 
 @testset "Copy sequence" begin