From a6fff2be46521d552c6633b3dbcd5ca659bc49e0 Mon Sep 17 00:00:00 2001 From: Jakob Nissen Date: Tue, 28 Jul 2020 13:17:50 +0200 Subject: [PATCH] Make BAM record layout match BAM specs --- src/bam/record.jl | 51 +++++++++++++++++++++++++---------------------- test/test_bam.jl | 32 +++++++++++++++++------------ 2 files changed, 46 insertions(+), 37 deletions(-) diff --git a/src/bam/record.jl b/src/bam/record.jl index a70bb17..e92eeae 100644 --- a/src/bam/record.jl +++ b/src/bam/record.jl @@ -11,18 +11,21 @@ mutable struct Record block_size::Int32 refid::Int32 pos::Int32 - bin_mq_nl::UInt32 - flag_nc::UInt32 + l_read_name::UInt8 + mapq::UInt8 + bin::UInt16 + n_cigar_op::UInt16 + flag::UInt16 l_seq::Int32 next_refid::Int32 next_pos::Int32 tlen::Int32 # variable length data data::Vector{UInt8} - reader::Reader + reader::Union{Reader, Nothing} function Record() - return new(0, 0, 0, 0, 0, 0, 0, 0, 0, UInt8[]) + return new(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, UInt8[]) end end @@ -49,8 +52,11 @@ function Base.:(==)(a::Record, b::Record) return a.block_size == b.block_size && a.refid == b.refid && a.pos == b.pos && - a.bin_mq_nl == b.bin_mq_nl && - a.flag_nc == b.flag_nc && + a.l_read_name == b.l_read_name && + a.mapq == b.mapq && + a.bin == b.bin && + a.n_cigar_op == b.n_cigar_op && + a.flag == b.flag && a.l_seq == b.l_seq && a.next_refid == b.next_refid && a.next_pos == b.next_pos && @@ -60,19 +66,13 @@ end function Base.copy(record::Record) copy = Record() - copy.block_size = record.block_size - copy.refid = record.refid - copy.pos = record.pos - copy.bin_mq_nl = record.bin_mq_nl - copy.flag_nc = record.flag_nc - copy.l_seq = record.l_seq - copy.next_refid = record.next_refid - copy.next_pos = record.next_pos - copy.tlen = record.tlen - copy.data = record.data[1:data_size(record)] - if isdefined(record, :reader) - copy.reader = record.reader + GC.@preserve copy record begin + dst_pointer = Ptr{UInt8}(pointer_from_objref(copy)) + src_pointer = Ptr{UInt8}(pointer_from_objref(record)) + unsafe_copyto!(dst_pointer, src_pointer, FIXED_FIELDS_BYTES) end + copy.data = record.data[1:data_size(record)] + copy.reader = record.reader return copy end @@ -80,8 +80,11 @@ function Base.empty!(record::Record) record.block_size = 0 record.refid = 0 record.pos = 0 - record.bin_mq_nl = 0 - record.flag_nc = 0 + record.l_read_name = 0 + record.mapq = 0 + record.bin = 0 + record.flag = 0 + record.n_cigar_op = 0 record.l_seq = 0 record.next_refid = 0 record.next_pos = 0 @@ -132,7 +135,7 @@ Get the bitwise flag of `record`. """ function flag(record::Record)::UInt16 checkfilled(record) - return UInt16(record.flag_nc >> 16) + return record.flag end function hasflag(record::Record) @@ -318,7 +321,7 @@ end Get the mapping quality of `record`. """ function mappingquality(record::Record)::UInt8 - return UInt8((record.bin_mq_nl >> 8) & 0xff) + return record.mapq end function hasmappingquality(record::Record) @@ -397,7 +400,7 @@ function extract_cigar_rle(data::Vector{UInt8}, offset, n) end function cigar_position(record::Record, checkCG::Bool = true)::Tuple{Int, Int} - cigaridx, nops = seqname_length(record) + 1, record.flag_nc & 0xFFFF + cigaridx, nops = seqname_length(record) + 1, record.n_cigar_op if !checkCG return cigaridx, nops end @@ -661,5 +664,5 @@ end # Return the length of the read name. function seqname_length(record::Record) - return record.bin_mq_nl & 0xff + return record.l_read_name end diff --git a/test/test_bam.jl b/test/test_bam.jl index 186d231..beb6d03 100644 --- a/test/test_bam.jl +++ b/test/test_bam.jl @@ -104,9 +104,12 @@ unsafe_copyto!(array, BAM.FIXED_FIELDS_BYTES + 1, record.data, 1, dsize) end new_record = convert(BAM.Record, array) - @test record.bin_mq_nl == new_record.bin_mq_nl + @test record.l_read_name == new_record.l_read_name + @test record.mapq == new_record.mapq + @test record.bin == new_record.bin @test record.block_size == new_record.block_size - @test record.flag_nc == new_record.flag_nc + @test record.flag == new_record.flag + @test record.n_cigar_op == new_record.n_cigar_op @test record.l_seq == new_record.l_seq @test record.next_refid == new_record.next_refid @test record.next_pos == new_record.next_pos @@ -162,18 +165,21 @@ if length(xs) != length(ys) return false end - for (x, y) in zip(xs, ys) + for (a, b) in zip(xs, ys) if !( - x.block_size == y.block_size && - x.refid == y.refid && - x.pos == y.pos && - x.bin_mq_nl == y.bin_mq_nl && - x.flag_nc == y.flag_nc && - x.l_seq == y.l_seq && - x.next_refid == y.next_refid && - x.next_pos == y.next_pos && - x.tlen == y.tlen && - x.data[1:BAM.data_size(x)] == y.data[1:BAM.data_size(y)]) + a.block_size == b.block_size && + a.refid == b.refid && + a.pos == b.pos && + a.l_read_name == b.l_read_name && + a.mapq == b.mapq && + a.bin == b.bin && + a.n_cigar_op == b.n_cigar_op && + a.flag == b.flag && + a.l_seq == b.l_seq && + a.next_refid == b.next_refid && + a.next_pos == b.next_pos && + a.tlen == b.tlen && + a.data[1:BAM.data_size(a)] == b.data[1:BAM.data_size(b)]) return false end end