Julia練習帳と質問

blast.jl

struct BlastRecord
    qseqid::String
    qseq::String
    sseqid::String
    sseq::String
    sstrand::String
    pident::Float64
    qlen::Int64
    length::Int64
    mismatch::Int64
    gapopen::Int64
    qstart::Int64
    qend::Int64
    sstart::Int64
    send::Int64
    gaps::Int64
    evalue::Float64
    bitscore::Float64
    btop::String
end

function generate_blast_records(blast_file::String)::Dict{String, BlastRecord}
    blast_db = Dict{String, BlastRecord}()
    open(blast_file, "r") do fh
        for line in eachline(fh)
            cols = split(line, "\t")
            qseqid = cols[1]
            qseq = cols[2]
            sseqid = cols[3]
            sseq = cols[4]
            sstrand = cols[5]
            pident = parse(Float64, cols[6])
            qlen = parse(Int, cols[7])
            length = parse(Int, cols[8])
            mismatch = parse(Int, cols[9])
            gapopen = parse(Int, cols[10])
            qstart = parse(Int, cols[11])
            qend = parse(Int, cols[12])
            sstart = parse(Int, cols[13])
            send = parse(Int, cols[14])
            gaps = parse(Int, cols[15])
            evalue = parse(Float64, cols[16])
            bitscore = parse(Float64, cols[17])
            btop = cols[18]

            record = BlastRecord(qseqid, qseq, sseqid, sseq, sstrand, pident,
                                 qlen, length, mismatch, gapopen, qstart, qend,
                                 sstart, send, gaps, evalue, bitscore, btop)
            blast_db[qseqid] = record
        end
    end
    return blast_db
end

@time blast_recs = generate_blast_records("OFF1_lib_R1_1.blast")
for key in keys(blast_recs)
    @show(key, blast_recs[key])
end

DataFrames.jl

## CSVでDataFrame型を作成し、これをクエリできるようになった

module Main

using ArgParse
using CSV
using DataFrames
using Printf

struct BlastRecord
    qseqid::String
    qseq::String
    sseqid::String
    sseq::String
    sstrand::String
    pident::Float64
    qlen::Int64
    length::Int64
    mismatch::Int64
    gapopen::Int64
    qstart::Int64
    qend::Int64
    sstart::Int64
    send::Int64
    gaps::Int64
    evalue::Float64
    bitscore::Float64
    btop::String
end

function generate_blast_records(blast_file::String)::Dict{String, BlastRecord}
    blast_db = Dict{String, BlastRecord}()
    open(blast_file, "r") do fh
        for line in eachline(fh)
            cols = split(line, "\t")
            qseqid = cols[1]
            qseq = cols[2]
            sseqid = cols[3]
            sseq = cols[4]
            sstrand = cols[5]
            pident = parse(Float64, cols[6])
            qlen = parse(Int, cols[7])
            length = parse(Int, cols[8])
            mismatch = parse(Int, cols[9])
            gapopen = parse(Int, cols[10])
            qstart = parse(Int, cols[11])
            qend = parse(Int, cols[12])
            sstart = parse(Int, cols[13])
            send = parse(Int, cols[14])
            gaps = parse(Int, cols[15])
            evalue = parse(Float64, cols[16])
            bitscore = parse(Float64, cols[17])
            btop = cols[18]
            
            record = BlastRecord(qseqid, qseq, sseqid, sseq, sstrand, pident,
                                 qlen, length, mismatch, gapopen, qstart, qend,
                                 sstart, send, gaps, evalue, bitscore, btop)
            blast_db[qseqid] = record
        end
    end
    return blast_db
end

function parse_command_line()::Dict{String, Any}
    s = ArgParseSettings(prog="blast_parser.jl", description="Parsing blast output file")
    @add_arg_table s begin
        "--blast"
        help = "Blast output"
        required = true
    end
    return parse_args(s)
end

function load_sample_sheet(file, delim)::DataFrame
    CSV.read(file, delim=delim)
end

function find_sample_by_id(sample_sheet::DataFrame, query::String)::DataFrameRow
    for sample in eachrow(sample_sheet)
        if (sample[:Sample_id] == query)
            return sample
        end
    end
    error("query \"$query\" not found in SampleSheet")
end

function collect_sample_id_by_enzyme(sample_sheet::DataFrame, query::String)::Vector{DataFrameRow}
    ret = Vector{DataFrameRow}()
    for sample in eachrow(sample_sheet)
        if (sample[:Enzyme] == query)
            push!(ret, sample)
        end
    end

    if length(ret) == 0
        error("no query $query found in SampleShet")
    else
        return ret
    end
end

function main()
    #args = parse_command_line()
    #blast_recs = Blast.generate_blast_records(args["blast"])
    #blast_recs = Blast.generate_blast_records("OFF001_WT_rBC.blast")
    
    file_path = "09052020_SampleSheet.tsv"
    sample_sheet = Main.load_sample_sheet(file_path, "\t")
    q = "v3_QC_1"
    ret = Main.find_sample_by_id(sample_sheet, q)
        
    ids = Main.collect_sample_id_by_enzyme(sample_sheet, "TagRFP control")
    @info "sample id:" ids[1:5]
end


## main
main()

end

• エラー処理の作法がいまいちわからない。例えば、なにか文字列処理をしていて結果を関数で返す場合、何もヒットがなかった場合が想定される場合は、関数はつねにUnion{T, Nothing}を返すべきなのかな？
• 例外はerrorかthrowがあるがどっちを使うべき？これって関数のbodyの最後にpythonでいうところのraise ValueError(...)的な感じで使えばいいのかな
• @infoマクロは便利だ。