Implemented an initial transcript merge stage

internal/model/model.go

@@ -2,7 +2,15 @@ package model
 
 // RawTranscript is a loaded input document before canonical normalization.
 type RawTranscript struct {
-	Source string `json:"source"`
+	Source   string       `json:"source"`
+	Segments []RawSegment `json:"segments"`
+}
+
+// RawSegment is the supported WhisperX segment subset.
+type RawSegment struct {
+	Start float64 `json:"start"`
+	End   float64 `json:"end"`
+	Text  string  `json:"text"`
 }
 
 // CanonicalTranscript is a per-speaker transcript in seriatim's internal model.