internal/adapters/transcribe/gateway.go

@@ -110,8 +110,9 @@ func (g *Gateway) Transcribe(ctx context.Context, path string) (domain.Transcrip
 }
 
 // verboseJSON is the subset of the verbose_json response used. Both backends
-// return it; they differ in segment granularity (iguana: whole seconds, berget:
-// sub-second) and in the language token, neither of which matters here.
+// return it; they differ in segment granularity (iguana: whole seconds and
+// sub-sentence pieces, berget: sub-second sentences), in the language token, and in
+// whether a segment's text starts with a space.
 type verboseJSON struct {
 	Text     string  `json:"text"`
 	Duration float64 `json:"duration"`
@@ -124,13 +125,21 @@ type verboseJSON struct {
 
 // parseVerbose normalises a verbose_json body. Whisper's final window can end past
 // the audio (9.0 s for 4.25 s of audio was observed), so segment ends are clamped
-// to the reported duration.
+// to the reported duration. Text is rebuilt from the segments, not taken from the
+// response's `text` field: see joinSegments.
 func parseVerbose(r io.Reader) (domain.Transcription, error) {
 	var v verboseJSON
 	if err := json.NewDecoder(r).Decode(&v); err != nil {
 		return domain.Transcription{}, fmt.Errorf("decode response: %w", err)
 	}
-	text := strings.TrimSpace(v.Text)
+	raw := make([]string, 0, len(v.Segments))
+	for _, s := range v.Segments {
+		raw = append(raw, s.Text)
+	}
+	text := joinSegments(raw)
+	if text == "" {
+		text = collapseSpace(v.Text)
+	}
 	if text == "" {
 		return domain.Transcription{}, ErrEmptyTranscript
 	}
@@ -145,3 +154,32 @@ func parseVerbose(r io.Reader) (domain.Transcription, error) {
 	}
 	return t, nil
 }
+
+// joinSegments rebuilds the running text from raw (untrimmed) segment texts.
+//
+// whisper.cpp wraps its output at ~60 characters and a wrap can land INSIDE a word
+// ("nyt" | "t godkännande"): a segment that begins with a space starts a new word,
+// one that does not continues the previous one, and the response's own `text`
+// field joins them with newlines, which breaks the word. So when any segment
+// carries a leading space the raw texts are concatenated as they are. Backends
+// that return whole sentences with no leading space (berget) are joined with a
+// space instead — concatenating those would glue sentences together.
+func joinSegments(raw []string) string {
+	leadingSpace := false
+	for _, s := range raw {
+		if strings.HasPrefix(s, " ") {
+			leadingSpace = true
+			break
+		}
+	}
+	sep := " "
+	if leadingSpace {
+		sep = ""
+	}
+	return collapseSpace(strings.Join(raw, sep))
+}
+
+// collapseSpace turns every run of whitespace (including newlines) into one space.
+func collapseSpace(s string) string {
+	return strings.Join(strings.Fields(s), " ")
+}