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

Switch to faster json library go-json and sonic

Browse Source 
Former-commit-id: ab9b4723f1dcf79fe5c073fff4d86f4f6969edfd
This commit is contained in:
Eric Coissac
2024-06-23 00:36:08 +02:00
parent 93f9dcb95f
commit c1f03cb1f6
9 changed files with 80 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
pkg/obialign/backtracking.go
View File

@ -1,12 +1,19 @@
package obialign
import "slices"
import (
log "github.com/sirupsen/logrus"
"slices"
)
func _Backtracking(pathMatrix []int, lseqA, lseqB int, path *[]int) []int {
needed := (lseqA + lseqB) * 2
(*path) = (*path)[:0]
cp := cap(*path)
(*path) = slices.Grow((*path), needed)
if cp < cap(*path) {
log.Infof("Resized path from %d to %d\n", cp, cap(*path))
}
p := cap(*path)
*path = (*path)[:p]
@ -88,7 +95,5 @@ func _Backtracking(pathMatrix []int, lseqA, lseqB int, path *[]int) []int {
(*path)[p] = 0
}
*path = (*path)[p:cap((*path))]
return *path
return (*path)[p:cap((*path))]
}

8
pkg/obialign/dnamatrix.go
View File

@ -2,6 +2,9 @@ package obialign
import (
"math"
"sync"
log "github.com/sirupsen/logrus"
)
var _FourBitsCount = []float64{
@ -28,6 +31,7 @@ var _InitializedDnaScore = false
var _NucPartMatch [32][32]float64
var _NucScorePartMatchMatch [100][100]int
var _NucScorePartMatchMismatch [100][100]int
var _InitDNAScoreMatrixMutex = &sync.Mutex{}
// _MatchRatio calculates the match ratio between two bytes.
//
@ -120,7 +124,11 @@ func _InitNucScorePartMatch() {
}
func _InitDNAScoreMatrix() {
_InitDNAScoreMatrixMutex.Lock()
defer _InitDNAScoreMatrixMutex.Unlock()
if !_InitializedDnaScore {
log.Info("Initializing the DNA Scoring matrix")
_InitNucPartMatch()
_InitNucScorePartMatch()
_InitializedDnaScore = true

40
pkg/obialign/pairedendalign.go
View File

@ -1,8 +1,6 @@
package obialign
import (
log "github.com/sirupsen/logrus"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obikmer"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq"
)
@ -319,7 +317,6 @@ func PELeftAlign(seqA, seqB *obiseq.BioSequence, gap, scale float64,
arena PEAlignArena) (int, []int) {
if !_InitializedDnaScore {
log.Println("Initializing the DNA Scoring matrix")
_InitDNAScoreMatrix()
}
@ -332,18 +329,17 @@ func PELeftAlign(seqA, seqB *obiseq.BioSequence, gap, scale float64,
&arena.pointer.scoreMatrix,
&arena.pointer.pathMatrix)
arena.pointer.path = _Backtracking(arena.pointer.pathMatrix,
path := _Backtracking(arena.pointer.pathMatrix,
seqA.Len(), seqB.Len(),
&arena.pointer.path)
return score, arena.pointer.path
return score, path
}
func PERightAlign(seqA, seqB *obiseq.BioSequence, gap, scale float64,
arena PEAlignArena) (int, []int) {
if !_InitializedDnaScore {
log.Println("Initializing the DNA Scoring matrix")
_InitDNAScoreMatrix()
}
@ -356,11 +352,11 @@ func PERightAlign(seqA, seqB *obiseq.BioSequence, gap, scale float64,
&arena.pointer.scoreMatrix,
&arena.pointer.pathMatrix)
arena.pointer.path = _Backtracking(arena.pointer.pathMatrix,
path := _Backtracking(arena.pointer.pathMatrix,
seqA.Len(), seqB.Len(),
&arena.pointer.path)
return score, arena.pointer.path
return score, path
}
func PEAlign(seqA, seqB *obiseq.BioSequence,
@ -373,8 +369,9 @@ func PEAlign(seqA, seqB *obiseq.BioSequence,
var rawSeqB, qualSeqB []byte
var extra5, extra3 int
var path []int
if !_InitializedDnaScore {
log.Println("Initializing the DNA Scoring matrix")
_InitDNAScoreMatrix()
}
@ -443,7 +440,7 @@ func PEAlign(seqA, seqB *obiseq.BioSequence,
&arena.pointer.pathMatrix)
}
arena.pointer.path = _Backtracking(arena.pointer.pathMatrix,
path = _Backtracking(arena.pointer.pathMatrix,
len(rawSeqA), len(rawSeqB),
&arena.pointer.path)
@ -474,15 +471,16 @@ func PEAlign(seqA, seqB *obiseq.BioSequence,
qualB := qualSeqB[i]
score += _NucScorePartMatchMatch[qualA][qualB]
}
arena.pointer.path = arena.pointer.path[:0]
arena.pointer.path = append(arena.pointer.path, 0, partLen)
path = arena.pointer.path[:0]
path = append(path, 0, partLen)
}
arena.pointer.path[0] += extra5
if arena.pointer.path[len(arena.pointer.path)-1] == 0 {
arena.pointer.path[len(arena.pointer.path)-2] += extra3
path[0] += extra5
if path[len(path)-1] == 0 {
path[len(path)-2] += extra3
} else {
arena.pointer.path = append(arena.pointer.path, extra3, 0)
path = append(path, extra3, 0)
}
} else {
//
@ -499,9 +497,9 @@ func PEAlign(seqA, seqB *obiseq.BioSequence,
&arena.pointer.scoreMatrix,
&arena.pointer.pathMatrix)
arena.pointer.path = _Backtracking(arena.pointer.pathMatrix,
path = _Backtracking(arena.pointer.pathMatrix,
len(rawSeqA), len(rawSeqB),
&arena.pointer.path)
&(arena.pointer.path))
scoreL := _FillMatrixPeLeftAlign(
rawSeqA, qualSeqA, rawSeqB, qualSeqB, gap, scale,
@ -509,12 +507,12 @@ func PEAlign(seqA, seqB *obiseq.BioSequence,
&arena.pointer.pathMatrix)
if scoreL > scoreR {
arena.pointer.path = _Backtracking(arena.pointer.pathMatrix,
path = _Backtracking(arena.pointer.pathMatrix,
len(rawSeqA), len(rawSeqB),
&arena.pointer.path)
&(arena.pointer.path))
}
}
return score, arena.pointer.path, fastCount, over, fastScore
return score, path, fastCount, over, fastScore
}

3
pkg/obiformats/dispatcher.go
View File

@ -1,13 +1,14 @@
package obiformats
import (
"encoding/json"
"fmt"
"os"
"path/filepath"
"strings"
"sync"
"github.com/goccy/go-json"
log "github.com/sirupsen/logrus"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiiter"

2
pkg/obiformats/json_writer.go
View File

@ -3,7 +3,7 @@ package obiformats
import (
"bufio"
"bytes"
"encoding/json"
"github.com/goccy/go-json"
"io"
"os"
"strconv"

2
pkg/obioptions/version.go
View File

@ -7,7 +7,7 @@ import (
// TODO: The version number is extracted from git. This induces that the version
// corresponds to the last commit, and not the one when the file will be
// commited
var _Commit = "fbdb2af"
var _Commit = "c94e79b"
var _Version = "Release 4.2.0"
// Version returns the version of the obitools package.

10
pkg/obiutils/goutils.go
View File

@ -3,12 +3,13 @@ package obiutils
import (
"bufio"
"bytes"
"encoding/json"
"io"
"os"
"reflect"
"sync"
"github.com/bytedance/sonic"
"github.com/barkimedes/go-deepcopy"
)
@ -363,9 +364,12 @@ func AtomicCounter(initial ...int) func() int {
return nextCounter
}
// JsonMarshalByteBuffer marshals an interface into JSON format.
//
// It takes a bytes.Buffer as a buffer and an interface{} as i.
// Returns an error.
func JsonMarshalByteBuffer(buffer *bytes.Buffer, i interface{}) error {
encoder := json.NewEncoder(buffer)
encoder.SetEscapeHTML(false)
encoder := sonic.ConfigDefault.NewEncoder(buffer)
err := encoder.Encode(i)
b := buffer.Bytes()
b = bytes.TrimRight(b, "\n")