1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
//! FASTQ reading with multi-line FASTQ support.
//!
//! # Behaviour
//!
//! * Since the quality scores can contain `@`, this parser always compares
//!   sequence and quality score lengths, internal `@` are ignored.
//! * Sequence and quality lines are *not* optional. The following is not a
//!   valid record: `@id\n+\n`. The minimal valid (though empty) record would
//!   be `@id\n\n+\n\n`.
//!
//! Writing multi-line FASTQ is not possible. This is on purpose, since
//! multi-line FASTQ is problematic and its use discouraged by many people.
//!
//! # Example
//!
//! ```rust
//! use seq_io::prelude::*;
//! use seq_io::fastq::multiline::Reader;
//!
//! # fn main() {
//! let seq = b"@id
//! SEQU
//! ENCE
//! +
//! II
//! @EI
//! III
//! ";
//!
//! let mut reader = Reader::new(&seq[..]);
//!
//! let rec = reader.next().unwrap().unwrap();
//!
//! assert_eq!(rec.id(), Ok("id"));
//! assert_eq!(rec.seq(), b"SEQU\nENCE");
//! assert_eq!(rec.full_seq().as_ref(), b"SEQUENCE");
//! assert_eq!(rec.full_qual().as_ref(), b"II@EIIII");
//! # }
//! ```
//!
//! **Note** that even if the second line of the quality string starts with a
//! `@`, it is recognized as internal quality score because it is assumed that
//! sequence and quality lengths are the same. If they were different, it may
//! confuse the parser and lead to weird errors.

use crate::core::{LinesParseKind, PositionStore, SearchPos};
use crate::LineSearchIter;
use serde::{Deserialize, Serialize};
use std::str;

impl_fastq_reader!(true, MultiRangeStore, ("fastq::multiline", "fastq"));

#[derive(Debug, Clone, Default, Serialize, Deserialize)]
pub struct MultiRangeStore {
    pos: [usize; 5],
    n_seq_lines: usize,
    n_qual_lines: usize,
}

impl PositionStore for MultiRangeStore {
    type SeqLinesType = LinesParseKind;
    type QualLinesType = LinesParseKind;

    #[inline]
    fn init(&mut self) {
        self.n_seq_lines = 0;
        self.n_qual_lines = 0;
    }

    #[inline]
    fn move_to_start(&mut self, search_pos: SearchPos, offset: usize) {
        for i in 0..search_pos as usize + 1 {
            self.pos[i] -= offset;
        }
    }

    #[inline]
    fn record_start(&self) -> usize {
        self.pos[0]
    }

    #[inline]
    fn set_record_start(&mut self, start: usize) {
        self.pos[0] = start;
    }

    #[inline]
    fn set_seq_start(&mut self, pos: usize) {
        self.pos[1] = pos;
        self.n_seq_lines += 1;
    }

    #[inline]
    fn add_seq_line_start(&mut self, _: usize) {
        self.n_seq_lines += 1;
    }

    #[inline]
    fn seq_start(&self) -> usize {
        self.pos[1]
    }

    #[inline]
    fn seq_starts(&self) -> &[usize] {
        &self.pos[1..2]
    }

    #[inline]
    fn set_sep_pos(&mut self, pos: usize, has_line: bool) {
        self.pos[2] = pos;
        if !has_line {
            self.n_seq_lines -= 1;
        }
    }

    #[inline]
    fn sep_pos(&self) -> usize {
        self.pos[2]
    }

    #[inline]
    fn set_qual_start(&mut self, pos: usize) {
        self.pos[3] = pos;
        self.n_qual_lines += 1;
    }

    #[inline]
    fn add_qual_line_start(&mut self, _: usize) {
        self.n_qual_lines += 1;
    }

    #[inline]
    fn qual_start(&self) -> usize {
        self.pos[3]
    }

    #[inline]
    fn qual_starts(&self) -> &[usize] {
        &self.pos[3..4]
    }

    #[inline]
    fn has_qual(&self) -> bool {
        self.qual_start() != 0
    }

    #[inline]
    fn record_end(&self) -> usize {
        self.pos[4]
    }

    #[inline]
    fn set_record_end(&mut self, pos: usize, has_line: bool) {
        self.pos[4] = pos;
        if !has_line && self.has_qual() {
            // one quality line too much may have been added
            self.n_qual_lines -= 1;
        }
    }

    #[inline]
    fn num_lines(&self) -> usize {
        1 + self.n_seq_lines + 1 + self.n_qual_lines
    }

    #[inline]
    fn line_offset(&self, search_pos: SearchPos, has_line: bool) -> usize {
        self.n_seq_lines + self.n_qual_lines + (search_pos >= SearchPos::SEP) as usize
            - !has_line as usize
    }

    #[inline]
    fn num_seq_lines(&self) -> usize {
        self.n_seq_lines
    }

    #[inline]
    fn num_qual_lines(&self) -> usize {
        self.n_qual_lines
    }

    #[inline]
    fn seq_lines<'s>(&'s self, buffer: &'s [u8]) -> LineSearchIter<'s> {
        LineSearchIter::new(
            &buffer[self.seq_start()..self.sep_pos()],
            self.n_seq_lines == 1,
        )
    }

    fn qual_lines<'s>(&'s self, buffer: &'s [u8]) -> LineSearchIter<'s> {
        // we need the newline at the end in order to correctly parse the lines
        // -> but records at EOF may not have a newline
        let end = if self.record_end() >= buffer.len() + 1 {
            debug_assert!(self.record_end() == buffer.len() + 1);
            buffer.len()
        } else {
            self.record_end()
        };
        LineSearchIter::new(&buffer[self.qual_start()..end], self.n_qual_lines == 1)
    }
}