package cannot be added in Jullia

[pdimens]

When trying to add package using Julia package manager (Arch linux 64), error is presented:

(v1.0) pkg> add Rifraf
 Resolving package versions...
ERROR: Unsatisfiable requirements detected for package Rifraf [48487f09]:
 Rifraf [48487f09] log:
 ├─possible versions are: 0.0.1 or uninstalled
 ├─restricted to versions * by an explicit requirement, leaving only versions 0.0.1
 └─restricted by julia compatibility requirements to versions: uninstalled — no versions left

[OscarAspelin]

Try using Julia 0.6 instead!

[sdwfrost]

I've been trying to get this to work with Julia 1.0+. I added Project.toml and Manifest.toml, changed Base.full to full in src/bandedarrays.jl, and temporarily moved the reference to TestSetExtensions, which also needs to be updated to 1.0. The package now installs, but throws an error when I try to run the algorithm (the FASTQ and FASTA reference files load in fine):

result=Rifraf.rifraf(sequences,phreds,reference)
ERROR: MethodError: no method matching rifraf(::Array{BioSequences.BioSequence{BioSequences.DNAAlphabet{4}},1}, ::Array{Array{Int8,1},1}, ::Array{BioSequences.BioSequence{BioSequences.DNAAlphabet{4}},1})
Closest candidates are:
  rifraf(::Array{BioSequences.BioSequence{BioSequences.DNAAlphabet{2}},1}, ::Array{Array{Int8,1},1}; kwargs...) at /home/simon/Programs/Rifraf.jl/src/model.jl:1281
Stacktrace:
 [1] top-level scope at none:0

Any idea why rifraf is expecting a sequence with an alphabet of size 2?

[murrellb]

Hi Simon!

A 1.0 port would be wonderful - I've not had the time yet!

From here:

DNAAlphabet{2} is an alphabet that uses two bits per base and limits to only unambiguous nucleotide symbols (ACGT in DNA and ACGU in RNA).

[sdwfrost]

So, does that mean that I can only use BioSequence{DNAAlphabet{2}} in Rifraf?

[sdwfrost]

I have a fair few ambiguous calls, and I can't simply replace them with N.

[murrellb]

I think so... I'd need to think about the amount of work required to get this to work with ambigs - it won't just be a type signature change somewhere, but it might not be too difficult.

Are your ambig calls completely 50/50? If one base is more likely, then you might get ok performance just using that base, with a suitably low quality score (which Rifraf respects).

We built this for PacBio CCS, which have no ambigs.