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

Changes to be committed:

Browse Source 
modified:   pkg/obiformats/ngsfilter_read.go
	modified:   pkg/obioptions/version.go
	modified:   pkg/obiutils/mimetypes.go
This commit is contained in:
Eric Coissac
2025-05-14 14:53:25 +02:00
parent f9324dd8f4
commit 3424d3057f
3 changed files with 33 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
pkg/obiformats/ngsfilter_read.go
View File

@ -89,7 +89,7 @@ func _parseMainNGSFilter(text string) (obingslibrary.PrimerPair, obingslibrary.T
} }
func NGSFilterCsvDetector(raw []byte, limit uint32) bool { func NGSFilterCsvDetector(raw []byte, limit uint32) bool {
r := csv.NewReader(bytes.NewReader(obiutils.DropLastLine(raw, limit))) r := csv.NewReader(bytes.NewReader(obiutils.DropLastLine(raw)))
r.Comma = ',' r.Comma = ','
r.ReuseRecord = true r.ReuseRecord = true
r.LazyQuotes = true r.LazyQuotes = true
@ -110,7 +110,6 @@ func NGSFilterCsvDetector(raw []byte, limit uint32) bool {
if err != nil { if err != nil {
return false return false
} }
if nfields == 0 { if nfields == 0 {
nfields = len(rec) nfields = len(rec)
} else if nfields != len(rec) { } else if nfields != len(rec) {
@ -133,16 +132,20 @@ func OBIMimeNGSFilterTypeGuesser(stream io.Reader) (*mimetype.MIME, io.Reader, e
return nil, nil, err return nil, nil, err
} }
buf = buf[:n]
obiutils.HasBOM(buf)
mimetype.Lookup("text/plain").Extend(NGSFilterCsvDetector, "text/ngsfilter-csv", ".csv") mimetype.Lookup("text/plain").Extend(NGSFilterCsvDetector, "text/ngsfilter-csv", ".csv")
// Detect the MIME type using the mimetype library // Detect the MIME type using the mimetype library
mimeType := mimetype.Detect(buf[:n]) mimeType := mimetype.Detect(buf)
if mimeType == nil { if mimeType == nil {
return nil, nil, err return nil, nil, err
} }
// Create a new reader based on the read data // Create a new reader based on the read data
newReader := io.Reader(bytes.NewReader(buf[:n])) newReader := io.Reader(bytes.NewReader(buf))
if err == nil { if err == nil {
newReader = io.MultiReader(newReader, stream) newReader = io.MultiReader(newReader, stream)

2
pkg/obioptions/version.go
View File

@ -8,7 +8,7 @@ import (
// corresponds to the last commit, and not the one when the file will be // corresponds to the last commit, and not the one when the file will be
// commited // commited
var _Commit = "f1b9ac4" var _Commit = "f9324dd"
var _Version = "Release 4.4.0" var _Version = "Release 4.4.0"
// Version returns the version of the obitools package. // Version returns the version of the obitools package.

29
pkg/obiutils/mimetypes.go
View File

@ -8,12 +8,33 @@ import (
"regexp" "regexp"
"github.com/gabriel-vasile/mimetype" "github.com/gabriel-vasile/mimetype"
log "github.com/sirupsen/logrus"
) )
func DropLastLine(b []byte, readLimit uint32) []byte { func HasBOM(data []byte) bool {
if readLimit == 0 || uint32(len(b)) < readLimit { switch {
return b case bytes.HasPrefix(data, []byte{0xEF, 0xBB, 0xBF}):
log.Infoln("BOM detected: UTF-8 (EF BB BF)")
return true
case bytes.HasPrefix(data, []byte{0xFE, 0xFF}):
log.Infoln("BOM detected: UTF-16 Big Endian (FE FF)")
return true
case bytes.HasPrefix(data, []byte{0xFF, 0xFE}):
log.Infoln("BOM detected: UTF-16 Little Endian (FF FE)")
return true
case bytes.HasPrefix(data, []byte{0x00, 0x00, 0xFE, 0xFF}):
log.Infoln("BOM detected: UTF-32 Big Endian (00 00 FE FF)")
return true
case bytes.HasPrefix(data, []byte{0xFF, 0xFE, 0x00, 0x00}):
log.Infoln("BOM detected: UTF-32 Little Endian (FF FE 00 00)")
return true
default:
log.Infoln("No BOM detected")
return false
} }
}
func DropLastLine(b []byte) []byte {
for i := len(b) - 1; i > 0; i-- { for i := len(b) - 1; i > 0; i-- {
if b[i] == '\n' { if b[i] == '\n' {
return b[:i] return b[:i]
@ -27,7 +48,7 @@ var __obimimetype_registred__ = false
func RegisterOBIMimeType() { func RegisterOBIMimeType() {
if !__obimimetype_registred__ { if !__obimimetype_registred__ {
csv := func(in []byte, limit uint32) bool { csv := func(in []byte, limit uint32) bool {
in = DropLastLine(in, limit) in = DropLastLine(in)
br := bytes.NewReader(in) br := bytes.NewReader(in)
r := csv.NewReader(br) r := csv.NewReader(br)