chore: add Apache License 2.0 and ignore build directory

src/captions/processor.go

@@ -1,120 +1,127 @@
-// Package captions turns audio packets into fixed or speech-aware captions.
+// Package captions turns audio packets into presentation-neutral caption events.
 package captions
 
 import (
-	"fmt"
-	"io"
+	"strings"
 
 	"tea.chunkbyte.com/kato/captioneer/src/audio"
 	"tea.chunkbyte.com/kato/captioneer/src/config"
-	"tea.chunkbyte.com/kato/captioneer/src/models"
 )
 
 type Processor struct {
-	settings  config.Settings
-	resources *models.Resources
-	output    io.Writer
+	settings    config.Settings
+	transcriber Transcriber
+	emit        func(Event) error
 
 	fixedSamples   []float32
 	previewSamples []float32
 	previewActive  bool
-	previewVisible bool
+	previewStart   int
 	nextPreviewAt  int
 	totalSamples   int
 }
 
-func New(settings config.Settings, resources *models.Resources, output io.Writer) *Processor {
+func New(settings config.Settings, transcriber Transcriber, emit func(Event) error) *Processor {
 	return &Processor{
 		settings:      settings,
-		resources:     resources,
-		output:        output,
+		transcriber:   transcriber,
+		emit:          emit,
 		nextPreviewAt: audio.DurationSamples(settings.ChunkDuration),
 	}
 }
 
-func (p *Processor) Accept(samples []float32) {
+func (p *Processor) Accept(samples []float32) error {
 	if len(samples) == 0 {
-		return
+		return nil
 	}
 	if p.settings.Mode == config.ModeFixed {
-		p.acceptFixed(samples)
-		return
+		return p.acceptFixed(samples)
 	}
-	p.acceptVAD(samples)
+	return p.acceptVAD(samples)
 }
 
-func (p *Processor) Flush() {
+func (p *Processor) Flush() error {
 	if p.settings.Mode == config.ModeFixed {
 		if len(p.fixedSamples) > 0 {
-			p.emitCommitted(p.totalSamples, p.totalSamples+len(p.fixedSamples), p.fixedSamples)
+			return p.emitFinal(p.totalSamples, p.totalSamples+len(p.fixedSamples), p.fixedSamples)
 		}
-		return
+		return nil
 	}
-	p.resources.FlushVAD()
-	p.drainVAD()
-	p.clearPreview()
+	p.transcriber.FlushVAD()
+	return p.drainVAD()
 }
 
-func (p *Processor) acceptFixed(samples []float32) {
+func (p *Processor) acceptFixed(samples []float32) error {
 	p.fixedSamples = append(p.fixedSamples, samples...)
 	chunkSize := audio.DurationSamples(p.settings.ChunkDuration)
 	for len(p.fixedSamples) >= chunkSize {
-		p.emitCommitted(p.totalSamples, p.totalSamples+chunkSize, p.fixedSamples[:chunkSize])
+		if err := p.emitFinal(p.totalSamples, p.totalSamples+chunkSize, p.fixedSamples[:chunkSize]); err != nil {
+			return err
+		}
 		p.fixedSamples = p.fixedSamples[chunkSize:]
 		p.totalSamples += chunkSize
 	}
+	return nil
 }
 
-func (p *Processor) acceptVAD(samples []float32) {
+func (p *Processor) acceptVAD(samples []float32) error {
 	p.totalSamples += len(samples)
-	p.resources.AcceptVAD(samples)
+	p.transcriber.AcceptVAD(samples)
 
 	if p.previewActive || audio.RMSDBFS(samples) >= p.settings.PreviewThresholdDB {
-		p.previewActive = true
+		if !p.previewActive {
+			p.previewActive = true
+			p.previewStart = p.totalSamples - len(samples)
+		}
 		p.previewSamples = append(p.previewSamples, samples...)
 		if len(p.previewSamples) >= p.nextPreviewAt {
-			p.emitPreview(p.previewSamples)
+			if err := p.emitProvisional(); err != nil {
+				return err
+			}
 			p.nextPreviewAt += audio.DurationSamples(p.settings.ChunkDuration)
 		}
 	}
-	p.drainVAD()
+	return p.drainVAD()
 }
 
-func (p *Processor) drainVAD() {
+func (p *Processor) drainVAD() error {
 	for {
-		segment, ok := p.resources.NextSpeechSegment()
+		start, samples, ok := p.transcriber.NextSpeechSegment()
 		if !ok {
-			return
+			return nil
+		}
+		if err := p.emitFinal(start, start+len(samples), samples); err != nil {
+			return err
 		}
-		p.clearPreview()
-		p.emitCommitted(segment.Start, segment.Start+len(segment.Samples), segment.Samples)
 		p.previewSamples = nil
 		p.previewActive = false
+		p.previewStart = 0
 		p.nextPreviewAt = audio.DurationSamples(p.settings.ChunkDuration)
 	}
 }
 
-func (p *Processor) emitPreview(samples []float32) {
-	text := p.resources.Decode(samples)
+func (p *Processor) emitProvisional() error {
+	text := strings.TrimSpace(p.transcriber.Decode(p.previewSamples))
 	if text == "" {
-		return
+		return nil
 	}
-	fmt.Fprintf(p.output, "\r\033[2K… %s", text)
-	p.previewVisible = true
+	return p.emit(Event{
+		Kind:      Provisional,
+		Text:      text,
+		StartedAt: audio.SamplesDuration(p.previewStart),
+		EndedAt:   audio.SamplesDuration(p.previewStart + len(p.previewSamples)),
+	})
 }
 
-func (p *Processor) emitCommitted(start, end int, samples []float32) {
-	text := p.resources.Decode(samples)
+func (p *Processor) emitFinal(start, end int, samples []float32) error {
+	text := strings.TrimSpace(p.transcriber.Decode(samples))
 	if text == "" {
-		return
-	}
-	p.clearPreview()
-	fmt.Fprintf(p.output, "[%s-%s] %s\n", audio.FormatTime(start), audio.FormatTime(end), text)
-}
-
-func (p *Processor) clearPreview() {
-	if p.previewVisible {
-		fmt.Fprint(p.output, "\r\033[2K")
-		p.previewVisible = false
+		return p.emit(Event{Kind: Hide})
 	}
+	return p.emit(Event{
+		Kind:      Final,
+		Text:      text,
+		StartedAt: audio.SamplesDuration(start),
+		EndedAt:   audio.SamplesDuration(end),
+	})
 }