Added some sync rails

Author: djthorpe

cmd/server/main.go

@@ -56,7 +56,7 @@ func main() {
 
 	// Create a whisper service
 	log.Println("Storing models at", dir)
-	whisper, err := whisper.New(dir)
+	whisper, err := whisper.New(dir, whisper.OptMaxConcurrent(1))
 	if err != nil {
 		log.Println(err)
 		os.Exit(-2)

go.sum

@@ -37,8 +37,8 @@ github.com/rivo/uniseg v0.4.7 h1:WUdvkW8uEhrYfLC4ZzdpI2ztxP1I582+49Oc5Mq64VQ=
 github.com/rivo/uniseg v0.4.7/go.mod h1:FN3SvrM+Zdj16jyLfmOkMNblXMcoc8DfTHruCPUcx88=
 github.com/stretchr/testify v1.9.0 h1:HtqpIVDClZ4nwg75+f6Lvsy/wHu+3BoSGCbBAcpTsTg=
 github.com/stretchr/testify v1.9.0/go.mod h1:r2ic/lqez/lEtzL7wO/rwa5dbSLXVDPFyf8C91i36aY=
-golang.org/x/exp v0.0.0-20240506185415-9bf2ced13842 h1:vr/HnozRka3pE4EsMEg1lgkXJkTFJCVUX+S/ZT6wYzM=
-golang.org/x/exp v0.0.0-20240506185415-9bf2ced13842/go.mod h1:XtvwrStGgqGPLc4cjQfWqZHG1YFdYs6swckp8vpsjnc=
+golang.org/x/exp v0.0.0-20240719175910-8a7402abbf56 h1:2dVuKD2vS7b0QIHQbpyTISPd0LeHDbnYEryqj5Q1ug8=
+golang.org/x/exp v0.0.0-20240719175910-8a7402abbf56/go.mod h1:M4RDyNAINzryxdtnbRXRL/OHtkFuWGRjvuhBJpk2IlY=
 golang.org/x/image v0.18.0 h1:jGzIakQa/ZXI1I0Fxvaa9W7yP25TqT6cHIHn+6CqvSQ=
 golang.org/x/image v0.18.0/go.mod h1:4yyo5vMFQjVjUcVk4jEQcU9MGy/rulF5WvUILseCM2E=
 golang.org/x/sys v0.22.0 h1:RI27ohtqKCnwULzJLqkv897zojh5/DwS/ENaMzUOaWI=

pkg/whisper/segmenter/segmenter.go

@@ -50,6 +50,11 @@ func (s *Segmenter) Close() error {
 //////////////////////////////////////////////////////////////////////////////
 // PUBLIC METHODS
 
+// TODO: segments are output through a callback, with the samples and a timestamp
+// TODO: we could do some basic silence and voice detection to segment to ensure
+// we don't overtax the CPU/GPU with silence and non-speech
+// TODO: We have hard-coded the sample format, sample rate and number of channels
+// here. We should make this configurable
 func (s *Segmenter) Decode(ctx context.Context) error {
 	mapFunc := func(stream int, params *ffmpeg.Par) (*ffmpeg.Par, error) {
 		if stream == s.reader.BestStream(media.AUDIO) {

pkg/whisper/task/context.go

@@ -5,6 +5,7 @@ import (
 	"encoding/json"
 	"fmt"
 	"path/filepath"
+	"sync"
 
 	// Packages
 	model "github.com/mutablelogic/go-whisper/pkg/whisper/model"
@@ -19,6 +20,8 @@ import (
 
 // Context is used for running the transcription or translation
 type Context struct {
+	sync.Mutex
+
 	model   string
 	whisper *whisper.Context
 
@@ -36,6 +39,9 @@ func New() *Context {
 
 // Init the context
 func (m *Context) Init(path string, model *model.Model, gpu int) error {
+	m.Lock()
+	defer m.Unlock()
+
 	// Check parameters
 	if model == nil {
 		return ErrBadParameter
@@ -143,25 +149,31 @@ func (task *Context) CanTranslate() bool {
 // TODO: We need a callback for segment progress.
 func (task *Context) Transcribe(ctx context.Context, samples []float32) error {
 	// Set the 'abort' function
-	task.params.SetAbortCallback(task.whisper, func() bool {
+	/*task.params.SetAbortCallback(task.whisper, func() bool {
 		select {
 		case <-ctx.Done():
 			return true
 		default:
 			return false
 		}
-	})
+	})*/
 
 	// Set the 'progress' function
-	task.params.SetProgressCallback(task.whisper, func(percent int) {
-		fmt.Printf("Progress: %v\n", percent)
-	})
+	//task.params.SetProgressCallback(task.whisper, func(percent int) {
+	//	fmt.Printf("Progress: %v\n", percent)
+	//})
 
 	// Perform the transcription
 	if err := whisper.Whisper_full(task.whisper, task.params, samples); err != nil {
 		return err
 	}
 
+	// Get segments
+	for i := 0; i < task.whisper.NumSegments(); i++ {
+		segment := task.whisper.Segment(i)
+		fmt.Printf("Segment: %v\n", segment.Text)
+	}
+
 	// Return success
 	return nil
 }

pkg/whisper/whisper.go

@@ -5,6 +5,7 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"runtime"
 	"strings"
 
 	// Packages
@@ -46,7 +47,7 @@ func New(path string, opt ...Opt) (*Whisper, error) {
 	var o opts
 
 	// Set options
-	o.MaxConcurrent = 1
+	o.MaxConcurrent = runtime.NumCPU()
 	for _, fn := range opt {
 		if err := fn(&o); err != nil {
 			return nil, err