obitools/obitools4
2
0
Fork 0
mirror of https://github.com/metabarcoding/obitools4.git synced 2025-06-29 16:20:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

Limit allocation during genbank parsing

Browse Source 
Former-commit-id: eee3c1fa7ffb79943109ee32dbf21e78bf11b14f
This commit is contained in:
Eric Coissac
2023-03-28 22:42:58 +07:00
parent 446ba06c63
commit e7b9ba3f30
1 changed files with 18 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
pkg/obiformats/genbank_read.go
View File

@ -6,6 +6,7 @@ import (
"io"
"os"
"path"
"regexp"
"strconv"
"strings"
@ -28,10 +29,12 @@ const (
inSequence gbstate = 4
)
var _seqlenght_rx = regexp.MustCompile(" +([0-9]+) bp")
func _ParseGenbankFile(source string,
input <-chan _FileChunk, out obiiter.IBioSequence,
chunck_order func() int) {
var err error
state := inHeader
for chunks := range input {
@ -40,6 +43,7 @@ func _ParseGenbankFile(source string,
sequences := make(obiseq.BioSequenceSlice, 0, 100)
sumlength:=0
id := ""
lseq := -1
scientificName := ""
defBytes := new(bytes.Buffer)
featBytes := new(bytes.Buffer)
@ -57,6 +61,18 @@ func _ParseGenbankFile(source string,
case strings.HasPrefix(line, "LOCUS "):
state = inEntry
id = strings.SplitN(line[12:], " ", 2)[0]
match_length := _seqlenght_rx.FindStringSubmatch(line)
if len(match_length) > 0 {
lseq,err = strconv.Atoi(match_length[1])
if err != nil {
lseq = -1
}
}
if lseq > 0 {
seqBytes = bytes.NewBuffer(obiseq.GetSlice(lseq + 20))
} else {
seqBytes = new(bytes.Buffer)
}
case strings.HasPrefix(line, "SOURCE "):
scientificName = strings.TrimSpace(line[12:])
case strings.HasPrefix(line, "DEFINITION "):
@ -92,9 +108,8 @@ func _ParseGenbankFile(source string,
sequences = make(obiseq.BioSequenceSlice, 0, 100)
sumlength = 0
}
defBytes = new(bytes.Buffer)
defBytes = bytes.NewBuffer(obiseq.GetSlice(200))
featBytes = new(bytes.Buffer)
seqBytes = new(bytes.Buffer)
nl = 0
sl = 0
default: