Merge pull request #1915 from majiayu000/fix-1842-feature-request-parallelize-au-0101-2335

feat: parallelize audio chunk transcription for improved performance

cmd/generate_changelog/incoming/1915.txt

@@ -0,0 +1,3 @@
+### PR [#1915](https://github.com/danielmiessler/Fabric/pull/1915) by [majiayu000](https://github.com/majiayu000): feat: parallelize audio chunk transcription for improved performance
+
+- Parallelize audio chunk transcription using goroutines for improved performance

internal/plugins/ai/openai/openai_audio.go

@@ -10,12 +10,20 @@ import (
 	"slices"
 	"sort"
 	"strings"
+	"sync"
 
 	debuglog "github.com/danielmiessler/fabric/internal/log"
 
 	openai "github.com/openai/openai-go"
 )
 
+// transcriptionResult holds the result of a single chunk transcription.
+type transcriptionResult struct {
+	index int
+	text  string
+	err   error
+}
+
 // MaxAudioFileSize defines the maximum allowed size for audio uploads (25MB).
 const MaxAudioFileSize int64 = 25 * 1024 * 1024
 
@@ -73,27 +81,56 @@ func (o *Client) TranscribeFile(ctx context.Context, filePath, model string, spl
 		files = []string{filePath}
 	}
 
-	var builder strings.Builder
+	resultsChan := make(chan transcriptionResult, len(files))
+	var wg sync.WaitGroup
+
 	for i, f := range files {
-		debuglog.Log("Using model %s to transcribe part %d (file name: %s)...\n", model, i+1, f)
+		wg.Add(1)
-		var chunk *os.File
+		go func(index int, filePath string) {
-		if chunk, err = os.Open(f); err != nil {
+			defer wg.Done()
-			return "", err
+			debuglog.Log("Using model %s to transcribe part %d (file name: %s)...\n", model, index+1, filePath)
+
+			chunk, openErr := os.Open(filePath)
+			if openErr != nil {
+				resultsChan <- transcriptionResult{index: index, err: openErr}
+				return
 			}
+			defer chunk.Close()
+
 			params := openai.AudioTranscriptionNewParams{
 				File:  chunk,
 				Model: openai.AudioModel(model),
 			}
-		var resp *openai.Transcription
+			resp, transcribeErr := o.ApiClient.Audio.Transcriptions.New(ctx, params)
-		resp, err = o.ApiClient.Audio.Transcriptions.New(ctx, params)
+			if transcribeErr != nil {
-		chunk.Close()
+				resultsChan <- transcriptionResult{index: index, err: transcribeErr}
-		if err != nil {
+				return
-			return "", err
 			}
+			resultsChan <- transcriptionResult{index: index, text: resp.Text}
+		}(i, f)
+	}
+
+	wg.Wait()
+	close(resultsChan)
+
+	results := make([]transcriptionResult, 0, len(files))
+	for result := range resultsChan {
+		if result.err != nil {
+			return "", result.err
+		}
+		results = append(results, result)
+	}
+
+	sort.Slice(results, func(i, j int) bool {
+		return results[i].index < results[j].index
+	})
+
+	var builder strings.Builder
+	for i, result := range results {
 		if i > 0 {
 			builder.WriteString(" ")
 		}
-		builder.WriteString(resp.Text)
+		builder.WriteString(result.text)
 	}
 
 	return builder.String(), nil