# BAM Record
# ==========

"""
    BAM.Record()

Create an unfilled BAM record.
"""
mutable struct Record
    # fixed-length fields (see BMA specs for the details)
    block_size::Int32
    refid::Int32
    pos::Int32
    bin_mq_nl::UInt32
    flag_nc::UInt32
    l_seq::Int32
    next_refid::Int32
    next_pos::Int32
    tlen::Int32
    # variable length data
    data::Vector{UInt8}
    reader::Reader

    function Record()
        return new(0, 0, 0, 0, 0, 0, 0, 0, 0, UInt8[])
    end
end

# the data size of fixed-length fields (block_size-tlen)
const FIXED_FIELDS_BYTES = 36

function Record(data::Vector{UInt8})
    return convert(Record, data)
end

function Base.convert(::Type{Record}, data::Vector{UInt8})
    length(data) < FIXED_FIELDS_BYTES && throw(ArgumentError("data too short"))
    record = Record()
    dst_pointer = Ptr{UInt8}(pointer_from_objref(record))
    unsafe_copyto!(dst_pointer, pointer(data), FIXED_FIELDS_BYTES)
    dsize = data_size(record)
    resize!(record.data, dsize)
    length(data) < dsize + FIXED_FIELDS_BYTES && throw(ArgumentError("data too short"))
    unsafe_copyto!(record.data, 1, data, FIXED_FIELDS_BYTES + 1, dsize)
    return record
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
    end
    return copy
end

function Base.show(io::IO, record::Record)
    print(io, summary(record), ':')
    if isfilled(record)
        println(io)
        println(io, "      template name: ", tempname(record))
        println(io, "               flag: ", flag(record))
        println(io, "       reference ID: ", refid(record))
        println(io, "           position: ", position(record))
        println(io, "    mapping quality: ", mappingquality(record))
        println(io, "              CIGAR: ", cigar(record))
        println(io, "  next reference ID: ", nextrefid(record))
        println(io, "      next position: ", nextposition(record))
        println(io, "    template length: ", templength(record))
        println(io, "           sequence: ", sequence(record))
        # TODO: pretty print base quality
        println(io, "       base quality: ", quality(record))
          print(io, "     auxiliary data:")
        for field in keys(auxdata(record))
            print(io, ' ', field, '=', record[field])
        end
    else
        print(io, " <not filled>")
    end
end

function Base.read!(reader::Reader, record::Record)
    return _read!(reader, record)
end


# Accessor Fuctions
# -----------------

"""
    flag(record::Record)::UInt16

Get the bitwise flag of `record`.
"""
function flag(record::Record)::UInt16
    checkfilled(record)
    return UInt16(record.flag_nc >> 16)
end

function hasflag(record::Record)
    return isfilled(record)
end

"""
    ismapped(record::Record)::Bool

Test if `record` is mapped.
"""
function ismapped(record::Record)::Bool
    return flag(record) & SAM.FLAG_UNMAP == 0
end

"""
    isprimary(record::Record)::Bool

Test if `record` is a primary line of the read.

This is equivalent to `flag(record) & 0x900 == 0`.
"""
function isprimary(record::Record)::Bool
    return flag(record) & 0x900 == 0
end

"""
    ispositivestrand(record::Record)::Bool

Test if `record` is aligned to the positive strand.

This is equivalent to `flag(record) & 0x10 == 0`.
"""
function ispositivestrand(record::Record)::Bool
    flag(record) & 0x10 == 0
end

"""
    refid(record::Record)::Int

Get the reference sequence ID of `record`.

The ID is 1-based (i.e. the first sequence is 1) and is 0 for a record without a mapping position.

See also: `BAM.rname`
"""
function refid(record::Record)::Int
    checkfilled(record)
    return record.refid + 1
end

function hasrefid(record::Record)
    return isfilled(record)
end

function checked_refid(record::Record)
    id = refid(record)
    if id == 0
        throw(ArgumentError("record is not mapped"))
    end
    if !isdefined(record, :reader)
        throw(ArgumentError("reader is not defined"))
    end
    return id
end

"""
    refname(record::Record)::String

Get the reference sequence name of `record`.

See also: `BAM.refid`
"""
function refname(record::Record)::String
    checkfilled(record)
    id = checked_refid(record)
    return record.reader.refseqnames[id]
end

"""
    reflen(record::Record)::Int

Get the length of the reference sequence this record applies to.
"""
function reflen(record::Record)::Int
    checkfilled(record)
    id = checked_refid(record)
    return record.reader.refseqlens[id]
end

function hasrefname(record::Record)
    return hasrefid(record)
end

"""
    position(record::Record)::Int

Get the 1-based leftmost mapping position of `record`.
"""
function position(record::Record)::Int
    checkfilled(record)
    return record.pos + 1
end

function hasposition(record::Record)
    return isfilled(record)
end

"""
    rightposition(record::Record)::Int

Get the 1-based rightmost mapping position of `record`.
"""
function rightposition(record::Record)::Int
    checkfilled(record)
    return Int32(position(record) + alignlength(record) - 1)
end

function hasrightposition(record::Record)
    return isfilled(record) && ismapped(record)
end

"""
    isnextmapped(record::Record)::Bool

Test if the mate/next read of `record` is mapped.
"""
function isnextmapped(record::Record)::Bool
    return isfilled(record) && (flag(record) & FLAG_MUNMAP == 0)
end

"""
    nextrefid(record::Record)::Int

Get the next/mate reference sequence ID of `record`.
"""
function nextrefid(record::Record)::Int
    checkfilled(record)
    return record.next_refid + 1
end

function hasnextrefid(record::Record)
    return isfilled(record)
end

"""
    nextrefname(record::Record)::String

Get the reference name of the mate/next read of `record`.
"""
function nextrefname(record::Record)::String
    checkfilled(record)
    id = nextrefid(record)
    if id == 0
        throw(ArgumentError("next record is not mapped"))
    elseif !isdefined(record, :reader)
        throw(ArgumentError("reader is not defined"))
    end
    return record.reader.refseqnames[id]
end

function hasnextrefname(record::Record)
    return isfilled(record) && isnextmapped(record)
end

"""
    nextposition(record::Record)::Int

Get the 1-based leftmost mapping position of the next/mate read of `record`.
"""
function nextposition(record::Record)::Int
    checkfilled(record)
    return record.next_pos + 1
end

function hasnextposition(record::Record)
    return isfilled(record)
end

"""
    mappingquality(record::Record)::UInt8

Get the mapping quality of `record`.
"""
function mappingquality(record::Record)::UInt8
    return UInt8((record.bin_mq_nl >> 8) & 0xff)
end

function hasmappingquality(record::Record)
    return isfilled(record)
end

"""
    n_cigar_op(record::Record, checkCG::Bool = true)

Return the number of operations in the CIGAR string of `record`.

Note that in the BAM specification, the field called `cigar` typically stores the cigar string of the record.
However, this is not always true, sometimes the true cigar is very long, and due to  some constraints of the BAM format, the actual cigar string is stored in an extra tag: `CG:B,I`, and the `cigar` field stores a pseudo-cigar string.

Calling this method with `checkCG` set to `true` (default) this method will always yield the number of operations in the true cigar string, because this is probably what you want, the vast majority of the time.

If you have a record that stores the true cigar in a `CG:B,I` tag, but you still want to get the number of operations in the `cigar` field of the BAM record, then set `checkCG` to `false`.
"""
function n_cigar_op(record::Record, checkCG::Bool = true)
    return cigar_position(record, checkCG)[2]
end

"""
    cigar(record::Record)::String

Get the CIGAR string of `record`.

Note that in the BAM specification, the field called `cigar` typically stores the cigar string of the record.
However, this is not always true, sometimes the true cigar is very long, and due to  some constraints of the BAM format, the actual cigar string is stored in an extra tag: `CG:B,I`, and the `cigar` field stores a pseudo-cigar string.

Calling this method with `checkCG` set to `true` (default) this method will always yield the true cigar string, because this is probably what you want the vast majority of the time.

If you have a record that stores the true cigar in a `CG:B,I` tag, but you still want to access the pseudo-cigar that is stored in the `cigar` field of the BAM record, then you can set checkCG to `false`.

See also `BAM.cigar_rle`.
"""
function cigar(record::Record, checkCG::Bool = true)::String
    buf = IOBuffer()
    for (op, len) in zip(cigar_rle(record, checkCG)...)
        print(buf, len, convert(Char, op))
    end
    return String(take!(buf))
end

"""
    cigar_rle(record::Record, checkCG::Bool = true)::Tuple{Vector{BioAlignments.Operation},Vector{Int}}

Get a run-length encoded tuple `(ops, lens)` of the CIGAR string in `record`.

Note that in the BAM specification, the field called `cigar` typically stores the cigar string of the record.
However, this is not always true, sometimes the true cigar is very long, and due to  some constraints of the BAM format, the actual cigar string is stored in an extra tag: `CG:B,I`, and the `cigar` field stores a pseudo-cigar string.

Calling this method with `checkCG` set to `true` (default) this method will always yield the true cigar string, because this is probably what you want the vast majority of the time.

If you have a record that stores the true cigar in a `CG:B,I` tag, but you still want to access the pseudo-cigar that is stored in the `cigar` field of the BAM record, then you can set checkCG to `false`.

See also `BAM.cigar`.
"""
function cigar_rle(record::Record, checkCG::Bool = true)::Tuple{Vector{BioAlignments.Operation},Vector{Int}}
    checkfilled(record)
    idx, nops = cigar_position(record, checkCG)
    ops, lens = extract_cigar_rle(record.data, idx, nops)
    return ops, lens
end

function extract_cigar_rle(data::Vector{UInt8}, offset, n)
    ops = Vector{BioAlignments.Operation}()
    lens = Vector{Int}()
    for i in offset:4:offset + (n - 1) * 4
        x = unsafe_load(Ptr{UInt32}(pointer(data, i)))
        op = BioAlignments.Operation(x & 0x0F)
        push!(ops, op)
        push!(lens, x >> 4)
    end
    return ops, lens
end

function cigar_position(record::Record, checkCG::Bool = true)::Tuple{Int, Int}
    cigaridx, nops = seqname_length(record) + 1, record.flag_nc & 0xFFFF
    if !checkCG
        return cigaridx, nops
    end
    if nops != 2
        return cigaridx, nops
    end
    x = unsafe_load(Ptr{UInt32}(pointer(record.data, cigaridx)))
    if x != UInt32(seqlength(record) << 4 | 4)
        return cigaridx, nops
    end
    start = auxdata_position(record)
    stop = data_size(record)
    tagidx = findauxtag(record.data, start, stop, UInt8('C'), UInt8('G'))
    if tagidx == 0
        return cigaridx, nops
    end
    # Tag exists, validate type is BI.
    typ = unsafe_load(Ptr{UInt16}(pointer(record.data, tagidx += 2)))
    if typ != (UInt16('I') << 8 | UInt16('B'))
        return cigaridx, nops
    end
    # If got this far, the CG tag is valid and contains the cigar.
    # Get the true n_cigar_ops, and return it and the idx of the first
    nops = UInt32(unsafe_load(Ptr{Int32}(pointer(record.data, tagidx += 2))))
    tagidx += 4
    return tagidx, nops
end

"""
    alignment(record::Record)::BioAlignments.Alignment

Get the alignment of `record`.
"""
function alignment(record::Record)::BioAlignments.Alignment
    checkfilled(record)
    if !ismapped(record)
        return BioAlignments.Alignment(BioAlignments.AlignmentAnchor[])
    end
    seqpos = 0
    refpos = position(record) - 1
    anchors = [BioAlignments.AlignmentAnchor(seqpos, refpos, BioAlignments.OP_START)]
    for (op, len) in zip(cigar_rle(record)...)
        if BioAlignments.ismatchop(op)
            seqpos += len
            refpos += len
        elseif BioAlignments.isinsertop(op)
            seqpos += len
        elseif BioAlignments.isdeleteop(op)
            refpos += len
        else
            error("operation $(op) is not supported")
        end
        push!(anchors, BioAlignments.AlignmentAnchor(seqpos, refpos, op))
    end
    return BioAlignments.Alignment(anchors)
end

function hasalignment(record::Record)
    return ismapped(record)
end

"""
    alignlength(record::Record)::Int

Get the alignment length of `record`.
"""
function alignlength(record::Record)::Int
    offset = seqname_length(record)
    length::Int = 0
    for i in offset + 1:4:offset + n_cigar_op(record, false) * 4
        x = unsafe_load(Ptr{UInt32}(pointer(record.data, i)))
        op = BioAlignments.Operation(x & 0x0F)
        if BioAlignments.ismatchop(op) || BioAlignments.isdeleteop(op)
            length += x >> 4
        end
    end
    return length
end

"""
    tempname(record::Record)::String

Get the query template name of `record`.
"""
function tempname(record::Record)::String
    checkfilled(record)
    # drop the last NUL character
    return unsafe_string(pointer(record.data), max(seqname_length(record) - 1, 0))
end

function hastempname(record::Record)
    return isfilled(record)
end

"""
    templength(record::Record)::Int

Get the template length of `record`.
"""
function templength(record::Record)::Int
    checkfilled(record)
    return record.tlen
end

function hastemplength(record::Record)
    return isfilled(record)
end

"""
    sequence(record::Record)::BioSequences.LongDNASeq

Get the segment sequence of `record`.
"""
function sequence(record::Record)::BioSequences.LongDNASeq
    checkfilled(record)
    seqlen = seqlength(record)
    data = Vector{UInt64}(undef, cld(seqlen, 16))
    src::Ptr{UInt64} = pointer(record.data, seqname_length(record) + n_cigar_op(record, false) * 4 + 1)
    for i in 1:lastindex(data)
        # copy data flipping high and low nybble
        x = unsafe_load(src, i)
        data[i] = (x & 0x0f0f0f0f0f0f0f0f) << 4 | (x & 0xf0f0f0f0f0f0f0f0) >> 4
    end
    return BioSequences.LongDNASeq(data, 1:seqlen, false)
end

function hassequence(record::Record)
    return isfilled(record)
end

"""
    seqlength(record::Record)::Int

Get the sequence length of `record`.
"""
function seqlength(record::Record)::Int
    checkfilled(record)
    return record.l_seq % Int
end

function hasseqlength(record::Record)
    return isfilled(record)
end

"""
    quality(record::Record)::Vector{UInt8}

Get the base quality of  `record`.
"""
function quality(record::Record)::Vector{UInt8}
    checkfilled(record)
    seqlen = seqlength(record)
    offset = seqname_length(record) + n_cigar_op(record, false) * 4 + cld(seqlen, 2)
    return [reinterpret(Int8, record.data[i+offset]) for i in 1:seqlen]
end

function hasquality(record::Record)
    return isfilled(record)
end

"""
    auxdata(record::Record)::BAM.AuxData

Get the auxiliary data of `record`.
"""
function auxdata(record::Record)
    checkfilled(record)
    return AuxData(record.data[auxdata_position(record):data_size(record)])
end

function hasauxdata(record::Record)
    return isfilled(record)
end

function Base.getindex(record::Record, tag::AbstractString)
    checkauxtag(tag)
    start = auxdata_position(record)
    stop = data_size(record)
    return getauxvalue(record.data, start, stop, UInt8(tag[1]), UInt8(tag[2]))
end

function Base.haskey(record::Record, tag::AbstractString)
    checkauxtag(tag)
    start = auxdata_position(record)
    stop = data_size(record)
    return findauxtag(record.data, start, stop, UInt8(tag[1]), UInt8(tag[2])) > 0
end

function Base.keys(record::Record)
    return collect(keys(auxdata(record)))
end

function Base.values(record::Record)
    return [record[key] for key in keys(record)]
end


# BioCore Methods
# -----------

function BioCore.isfilled(record::Record)
    return record.block_size != 0
end

function BioCore.seqname(record::Record)
    return tempname(record)
end

function BioCore.hasseqname(record::Record)
    return hastempname(record)
end

function BioCore.sequence(record::Record)
    return sequence(record)
end

function BioCore.hassequence(record::Record)
    return hassequence(record)
end

function BioCore.leftposition(record::Record)
    return position(record)
end

function BioCore.hasleftposition(record::Record)
    return hasposition(record)
end

function BioCore.rightposition(record::Record)
    return rightposition(record)
end

function BioCore.hasrightposition(record::Record)
    return hasrightposition(record)
end


# Helper Functions
# ----------------

# Return the size of the `.data` field.
function data_size(record::Record)
    if isfilled(record)
        return record.block_size - FIXED_FIELDS_BYTES + sizeof(record.block_size)
    end

    return 0

end

function checkfilled(record::Record)
    if !isfilled(record)
        throw(ArgumentError("unfilled BAM record"))
    end
end

function auxdata_position(record::Record)
    seqlen = seqlength(record)
    return seqname_length(record) + n_cigar_op(record, false) * 4 + cld(seqlen, 2) + seqlen + 1
end

# Return the length of the read name.
function seqname_length(record::Record)
    return record.bin_mq_nl & 0xff
end