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

Patch an algorithm error on th Annotation Classifier

Browse Source
This commit is contained in:
Eric Coissac
2022-08-21 14:45:53 +02:00
parent f1e3a68dc4
commit 10663bf79c
1 changed files with 10 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
pkg/obiseq/class.go
View File

@ -3,9 +3,10 @@ package obiseq
import ( import (
"fmt" "fmt"
"hash/crc32" "hash/crc32"
log "github.com/sirupsen/logrus"
"strconv" "strconv"
"sync" "sync"
log "github.com/sirupsen/logrus"
) )
type BioSequenceClassifier struct { type BioSequenceClassifier struct {
@ -15,6 +16,8 @@ type BioSequenceClassifier struct {
Clone func() *BioSequenceClassifier Clone func() *BioSequenceClassifier
} }
// It creates a classifier that returns the value of the annotation key as an integer. If the
// annotation key is not present, it returns the integer value of the string na
func AnnotationClassifier(key string, na string) *BioSequenceClassifier { func AnnotationClassifier(key string, na string) *BioSequenceClassifier {
encode := make(map[string]int, 1000) encode := make(map[string]int, 1000)
decode := make([]string, 0, 1000) decode := make([]string, 0, 1000)
@ -22,10 +25,10 @@ func AnnotationClassifier(key string, na string) *BioSequenceClassifier {
maxcode := 0 maxcode := 0
code := func(sequence *BioSequence) int { code := func(sequence *BioSequence) int {
var val string var val = na
var ok bool
if sequence.HasAnnotation() { if sequence.HasAnnotation() {
value, ok := sequence.Annotations()[key] value, ok := sequence.Annotations()[key]
if ok { if ok {
switch value := value.(type) { switch value := value.(type) {
case string: case string:
@ -35,7 +38,6 @@ func AnnotationClassifier(key string, na string) *BioSequenceClassifier {
} }
} }
} }
val = na
locke.Lock() locke.Lock()
defer locke.Unlock() defer locke.Unlock()
@ -80,6 +82,8 @@ func AnnotationClassifier(key string, na string) *BioSequenceClassifier {
return &c return &c
} }
// It takes a predicate function and returns a classifier that returns 1 if the predicate is true and 0
// otherwise
func PredicateClassifier(predicate SequencePredicate) *BioSequenceClassifier { func PredicateClassifier(predicate SequencePredicate) *BioSequenceClassifier {
code := func(sequence *BioSequence) int { code := func(sequence *BioSequence) int {
if predicate(sequence) { if predicate(sequence) {
@ -112,7 +116,6 @@ func PredicateClassifier(predicate SequencePredicate) *BioSequenceClassifier {
} }
// Builds a classifier function based on CRC32 of the sequence // Builds a classifier function based on CRC32 of the sequence
//
func HashClassifier(size int) *BioSequenceClassifier { func HashClassifier(size int) *BioSequenceClassifier {
code := func(sequence *BioSequence) int { code := func(sequence *BioSequence) int {
return int(crc32.ChecksumIEEE(sequence.Sequence()) % uint32(size)) return int(crc32.ChecksumIEEE(sequence.Sequence()) % uint32(size))
@ -135,7 +138,6 @@ func HashClassifier(size int) *BioSequenceClassifier {
} }
// Builds a classifier function based on the sequence // Builds a classifier function based on the sequence
//
func SequenceClassifier() *BioSequenceClassifier { func SequenceClassifier() *BioSequenceClassifier {
encode := make(map[string]int, 1000) encode := make(map[string]int, 1000)
decode := make([]string, 0, 1000) decode := make([]string, 0, 1000)
@ -190,6 +192,8 @@ func SequenceClassifier() *BioSequenceClassifier {
return &c return &c
} }
// It returns a classifier that assigns each sequence to a different class, cycling through the classes
// in order
func RotateClassifier(size int) *BioSequenceClassifier { func RotateClassifier(size int) *BioSequenceClassifier {
n := 0 n := 0
lock := sync.Mutex{} lock := sync.Mutex{}