Julia 1.0 Compatibility #41
Conversation
|
Thanks for taking the time to do this @dcjones! I'll merge when the other two PRs are in 👍 |
REQUIRE
Outdated
| IntervalTrees | ||
| Twiddle | ||
| DataStructures | ||
| Nullables |
There was a problem hiding this comment.
It is okay to have this now but I think that in general we should not depend on deprecated/removed features. We may need a second round to review our APIs.
There was a problem hiding this comment.
Nullables are only used to store some optional fields in the FASTA and 2bit parsers, I can change this to the new Union{T, Nothing} style this evening whilst I'm waiting on the Twiddle updates to make it into METADATA.
|
@dcjones To take this furthur and get CI passes I'm currently waiting on YAML, the version 1.0 changes PR is waiting on a Codecs release, which is here: JuliaLang/METADATA.jl#16489, but as it's not on attobot it's not moving very far. Can you head to the PR and usher them along a bit, or turn on attobot for Codecs and YAML? :) |
Codecov Report
@@ Coverage Diff @@
## version-1.0 #41 +/- ##
==============================================
Coverage ? 94.07%
==============================================
Files ? 40
Lines ? 1925
Branches ? 0
==============================================
Hits ? 1811
Misses ? 114
Partials ? 0
Continue to review full report at Codecov.
|
|
@dcjones Ok I've made a series of commits trying to get tests to pass. I've got them passing for 1.0, but curiously not for 0.7. What's making me tear my hair out is julia complains "ERROR: LoadError: LoadError: LoadError: syntax: invalid variable expression in "where"". Whilst pointing to "in expression starting at /home/travis/build/BioJulia/BioSequences.jl/src/bioseq/constructors.jl:65". At which is a constructor, which does not make use of the "where" expression anywhere in it's definition. What do you think? |
|
So I had a go and could reproduce your issue @benjward. A MWE without imports that works on v0.7 but not on v1.0 is: abstract type Alphabet end
abstract type Sequence end
struct DNAAlphabet{n} <: Alphabet end
mutable struct BioSequence{A<:Alphabet} <: Sequence
data::Vector{UInt64} # encoded character sequence data
part::UnitRange{Int} # interval within `data` defining the (sub)sequence
shared::Bool # true if and only if `data` is shared between sequences
function BioSequence{A}(data::Vector{UInt64},
part::UnitRange{Int},
shared::Bool) where A
return new(data, part, shared)
end
end
function BioSequence{DNAAlphabet{4}}(seq::BioSequence{DNAAlphabet{2}}) # errors here
newseq = BioSequence{vim}(length(seq))
#for (i, x) in enumerate(seq)
# unsafe_setindex!(newseq, x, i)
#end
return newseq
endIt seems that if you replace the offending line with function BioSequence{T}(seq::BioSequence{DNAAlphabet{2}}) where {T<:DNAAlphabet{4}}then it seems to work, which is strange. I made a PR to implement this across this file and tests seem to pass on both v0.7 and v1.0. Feel free to use it if you like, but I'm wondering if this is a Julia issue? |
|
Thanks @jgreener64! You're option made me think of another alternative, which also passes both of 0.7 and 1.0! abstract type Alphabet end
abstract type Sequence end
struct DNAAlphabet{n} <: Alphabet end
mutable struct BioSequence{A<:Alphabet} <: Sequence
data::Vector{UInt64} # encoded character sequence data
part::UnitRange{Int} # interval within `data` defining the (sub)sequence
shared::Bool # true if and only if `data` is shared between sequences
function BioSequence{A}(data::Vector{UInt64},
part::UnitRange{Int},
shared::Bool) where A
return new(data, part, shared)
end
end
function (::Type{BioSequence{DNAAlphabet{4}}})(seq::BioSequence{DNAAlphabet{2}})
newseq = BioSequence{DNAAlphabet{4}}(length(seq))
for (i, x) in enumerate(seq)
unsafe_setindex!(newseq, x, i)
end
return newseq
end |
|
Merging into the |
This was an ordeal, and ended up being a huge patch, but should bring sequences up to date with julia 1.0 changes.
Needs a couple other PRs to work: