| line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
|
1
|
|
|
|
|
|
|
/* FASTA/FASTQ parser using kseq.h */ |
|
2
|
|
|
|
|
|
|
#include "EXTERN.h" |
|
3
|
|
|
|
|
|
|
#include "perl.h" |
|
4
|
|
|
|
|
|
|
#include "XSUB.h" |
|
5
|
|
|
|
|
|
|
|
|
6
|
|
|
|
|
|
|
#include |
|
7
|
|
|
|
|
|
|
#include "kseq.h" |
|
8
|
|
|
|
|
|
|
|
|
9
|
|
|
|
|
|
|
// Initialize kseq |
|
10
|
36
|
100
|
|
|
|
|
KSEQ_INIT(gzFile, gzread) |
|
|
5
|
50
|
|
|
|
|
|
|
|
1
|
50
|
|
|
|
|
|
|
|
1
|
0
|
|
|
|
|
|
|
|
5
|
50
|
|
|
|
|
|
|
|
17
|
100
|
|
|
|
|
|
|
|
5
|
50
|
|
|
|
|
|
|
|
1
|
100
|
|
|
|
|
|
|
|
1
|
50
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
11
|
|
|
|
|
|
|
|
|
12
|
|
|
|
|
|
|
// Helper function to convert a kseq_t record to a Perl hash reference |
|
13
|
2
|
|
|
|
|
|
static SV* kseq_to_hash(pTHX_ kseq_t *seq) { |
|
14
|
2
|
|
|
|
|
|
HV* hash = newHV(); |
|
15
|
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
// Add name, always present |
|
17
|
2
|
|
|
|
|
|
hv_store(hash, "name", 4, newSVpvn(seq->name.s, seq->name.l), 0); |
|
18
|
|
|
|
|
|
|
|
|
19
|
|
|
|
|
|
|
// Add sequence, always present |
|
20
|
2
|
|
|
|
|
|
hv_store(hash, "seq", 3, newSVpvn(seq->seq.s, seq->seq.l), 0); |
|
21
|
|
|
|
|
|
|
|
|
22
|
|
|
|
|
|
|
// Add comment if present |
|
23
|
2
|
50
|
|
|
|
|
if (seq->comment.l) |
|
24
|
2
|
|
|
|
|
|
hv_store(hash, "comment", 7, newSVpvn(seq->comment.s, seq->comment.l), 0); |
|
25
|
|
|
|
|
|
|
|
|
26
|
|
|
|
|
|
|
// Add quality if present |
|
27
|
2
|
50
|
|
|
|
|
if (seq->qual.l) |
|
28
|
0
|
|
|
|
|
|
hv_store(hash, "qual", 4, newSVpvn(seq->qual.s, seq->qual.l), 0); |
|
29
|
|
|
|
|
|
|
|
|
30
|
2
|
|
|
|
|
|
return newRV_noinc((SV*)hash); |
|
31
|
|
|
|
|
|
|
} |
|
32
|
|
|
|
|
|
|
|
|
33
|
|
|
|
|
|
|
MODULE = FASTX::XS PACKAGE = FASTX::XS |
|
34
|
|
|
|
|
|
|
PROTOTYPES: DISABLE |
|
35
|
|
|
|
|
|
|
|
|
36
|
|
|
|
|
|
|
SV* |
|
37
|
|
|
|
|
|
|
_xs_new(class, filename) |
|
38
|
|
|
|
|
|
|
char* class |
|
39
|
|
|
|
|
|
|
char* filename |
|
40
|
|
|
|
|
|
|
CODE: |
|
41
|
|
|
|
|
|
|
gzFile fp; |
|
42
|
|
|
|
|
|
|
kseq_t *seq; |
|
43
|
|
|
|
|
|
|
|
|
44
|
|
|
|
|
|
|
// Open the file |
|
45
|
1
|
|
|
|
|
|
fp = gzopen(filename, "r"); |
|
46
|
1
|
50
|
|
|
|
|
if (fp == NULL) |
|
47
|
0
|
|
|
|
|
|
croak("Failed to open file: %s", filename); |
|
48
|
|
|
|
|
|
|
|
|
49
|
|
|
|
|
|
|
// Initialize kseq |
|
50
|
1
|
|
|
|
|
|
seq = kseq_init(fp); |
|
51
|
1
|
50
|
|
|
|
|
if (seq == NULL) { |
|
52
|
0
|
|
|
|
|
|
gzclose(fp); |
|
53
|
0
|
|
|
|
|
|
croak("Failed to initialize sequence parser"); |
|
54
|
|
|
|
|
|
|
} |
|
55
|
|
|
|
|
|
|
|
|
56
|
|
|
|
|
|
|
// Create a hash to store our object data |
|
57
|
1
|
|
|
|
|
|
HV* self = newHV(); |
|
58
|
|
|
|
|
|
|
|
|
59
|
|
|
|
|
|
|
// Store the file pointer and seq object as an IV |
|
60
|
1
|
|
|
|
|
|
hv_store(self, "_fp", 3, newSViv(PTR2IV(fp)), 0); |
|
61
|
1
|
|
|
|
|
|
hv_store(self, "_seq", 4, newSViv(PTR2IV(seq)), 0); |
|
62
|
|
|
|
|
|
|
|
|
63
|
|
|
|
|
|
|
// Bless and return |
|
64
|
1
|
|
|
|
|
|
RETVAL = sv_bless(newRV_noinc((SV*)self), gv_stashpv(class, 0)); |
|
65
|
|
|
|
|
|
|
OUTPUT: |
|
66
|
|
|
|
|
|
|
RETVAL |
|
67
|
|
|
|
|
|
|
|
|
68
|
|
|
|
|
|
|
SV* |
|
69
|
|
|
|
|
|
|
next_seq(self) |
|
70
|
|
|
|
|
|
|
SV* self |
|
71
|
|
|
|
|
|
|
CODE: |
|
72
|
|
|
|
|
|
|
HV* hash; |
|
73
|
|
|
|
|
|
|
SV** fp_sv; |
|
74
|
|
|
|
|
|
|
SV** seq_sv; |
|
75
|
|
|
|
|
|
|
gzFile fp; |
|
76
|
|
|
|
|
|
|
kseq_t *seq; |
|
77
|
|
|
|
|
|
|
int ret; |
|
78
|
|
|
|
|
|
|
|
|
79
|
|
|
|
|
|
|
// Get the hash |
|
80
|
3
|
50
|
|
|
|
|
if (!SvROK(self) || SvTYPE(SvRV(self)) != SVt_PVHV) |
|
|
|
50
|
|
|
|
|
|
|
81
|
0
|
|
|
|
|
|
croak("Not a blessed hash reference"); |
|
82
|
3
|
|
|
|
|
|
hash = (HV*)SvRV(self); |
|
83
|
|
|
|
|
|
|
|
|
84
|
|
|
|
|
|
|
// Get the file pointer and seq object |
|
85
|
3
|
|
|
|
|
|
fp_sv = hv_fetch(hash, "_fp", 3, 0); |
|
86
|
3
|
|
|
|
|
|
seq_sv = hv_fetch(hash, "_seq", 4, 0); |
|
87
|
|
|
|
|
|
|
|
|
88
|
3
|
50
|
|
|
|
|
if (!fp_sv || !seq_sv) |
|
|
|
50
|
|
|
|
|
|
|
89
|
0
|
|
|
|
|
|
croak("Invalid object"); |
|
90
|
|
|
|
|
|
|
|
|
91
|
|
|
|
|
|
|
// Check if pointers are valid (not already freed) |
|
92
|
3
|
50
|
|
|
|
|
if (!SvOK(*fp_sv) || SvIV(*fp_sv) == 0) |
|
|
|
50
|
|
|
|
|
|
|
93
|
0
|
|
|
|
|
|
croak("File pointer has been freed"); |
|
94
|
3
|
50
|
|
|
|
|
if (!SvOK(*seq_sv) || SvIV(*seq_sv) == 0) |
|
|
|
50
|
|
|
|
|
|
|
95
|
0
|
|
|
|
|
|
croak("Sequence parser has been freed"); |
|
96
|
|
|
|
|
|
|
|
|
97
|
3
|
|
|
|
|
|
fp = INT2PTR(gzFile, SvIV(*fp_sv)); |
|
98
|
3
|
|
|
|
|
|
seq = INT2PTR(kseq_t*, SvIV(*seq_sv)); |
|
99
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
// Read next sequence |
|
101
|
3
|
|
|
|
|
|
ret = kseq_read(seq); |
|
102
|
|
|
|
|
|
|
|
|
103
|
3
|
100
|
|
|
|
|
if (ret < 0) { |
|
104
|
|
|
|
|
|
|
// EOF or error |
|
105
|
1
|
|
|
|
|
|
RETVAL = &PL_sv_undef; |
|
106
|
|
|
|
|
|
|
} else { |
|
107
|
|
|
|
|
|
|
// Convert to hash and return |
|
108
|
2
|
|
|
|
|
|
RETVAL = kseq_to_hash(aTHX_ seq); |
|
109
|
|
|
|
|
|
|
} |
|
110
|
|
|
|
|
|
|
OUTPUT: |
|
111
|
|
|
|
|
|
|
RETVAL |
|
112
|
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
void |
|
114
|
|
|
|
|
|
|
DESTROY(self) |
|
115
|
|
|
|
|
|
|
SV* self |
|
116
|
|
|
|
|
|
|
CODE: |
|
117
|
|
|
|
|
|
|
HV* hash; |
|
118
|
|
|
|
|
|
|
SV** fp_sv; |
|
119
|
|
|
|
|
|
|
SV** seq_sv; |
|
120
|
|
|
|
|
|
|
gzFile fp; |
|
121
|
|
|
|
|
|
|
kseq_t *seq; |
|
122
|
|
|
|
|
|
|
|
|
123
|
|
|
|
|
|
|
// Get the hash |
|
124
|
1
|
50
|
|
|
|
|
if (!SvROK(self) || SvTYPE(SvRV(self)) != SVt_PVHV) |
|
|
|
50
|
|
|
|
|
|
|
125
|
0
|
|
|
|
|
|
return; |
|
126
|
1
|
|
|
|
|
|
hash = (HV*)SvRV(self); |
|
127
|
|
|
|
|
|
|
|
|
128
|
|
|
|
|
|
|
// Get the file pointer and seq object |
|
129
|
1
|
|
|
|
|
|
fp_sv = hv_fetch(hash, "_fp", 3, 0); |
|
130
|
1
|
|
|
|
|
|
seq_sv = hv_fetch(hash, "_seq", 4, 0); |
|
131
|
|
|
|
|
|
|
|
|
132
|
1
|
50
|
|
|
|
|
if (!fp_sv || !seq_sv) |
|
|
|
50
|
|
|
|
|
|
|
133
|
0
|
|
|
|
|
|
return; |
|
134
|
|
|
|
|
|
|
|
|
135
|
|
|
|
|
|
|
// Check if pointers are valid before cleanup |
|
136
|
1
|
50
|
|
|
|
|
if (SvOK(*fp_sv) && SvIV(*fp_sv) != 0) { |
|
|
|
50
|
|
|
|
|
|
|
137
|
1
|
|
|
|
|
|
fp = INT2PTR(gzFile, SvIV(*fp_sv)); |
|
138
|
1
|
|
|
|
|
|
gzclose(fp); |
|
139
|
|
|
|
|
|
|
// Mark as cleaned up |
|
140
|
1
|
|
|
|
|
|
sv_setiv(*fp_sv, 0); |
|
141
|
|
|
|
|
|
|
} |
|
142
|
|
|
|
|
|
|
|
|
143
|
1
|
50
|
|
|
|
|
if (SvOK(*seq_sv) && SvIV(*seq_sv) != 0) { |
|
|
|
50
|
|
|
|
|
|
|
144
|
1
|
|
|
|
|
|
seq = INT2PTR(kseq_t*, SvIV(*seq_sv)); |
|
145
|
1
|
|
|
|
|
|
kseq_destroy(seq); |
|
146
|
|
|
|
|
|
|
// Mark as cleaned up |
|
147
|
1
|
|
|
|
|
|
sv_setiv(*seq_sv, 0); |
|
148
|
|
|
|
|
|
|
} |