add_pause_between_segments

pause.py

@@ -7,13 +7,14 @@ def annotate_pauses(session_id, threshold, base_dir="session_data"):
     json_file = os.path.join(session_dir, f"{session_id}_transcriptionCW.json")
     
     if not os.path.exists(json_file):
-        print(f"Error: could not finf {json_file}")
+        print(f"Error: could not find {json_file}")
         return
     
     with open(json_file, "r", encoding="utf-8") as f:
         data = json.load(f)
     
     segments = data.get("segments", [])
+    
     for segment in segments:
         words = segment.get("words", [])
         if "pauses" in segment:
@@ -34,11 +35,52 @@ def annotate_pauses(session_id, threshold, base_dir="session_data"):
                     pauses.append(pause_info)
         segment["pauses"] = pauses
     
+    new_segments = []
+    
+    for i, segment in enumerate(segments):
+
+        new_segments.append(segment)
+        
+        if i < len(segments) - 1:
+            next_segment = segments[i + 1]
+            
+            current_words = segment.get("words", [])
+            next_words = next_segment.get("words", [])
+            
+            if current_words and next_words:
+                last_word_end = current_words[-1]["end"]
+                next_word_start = next_words[0]["start"]
+                gap = next_word_start - last_word_end
+                
+                if gap > threshold:
+                    pause_segment = {
+                        "start": round(last_word_end, 3),
+                        "end": round(next_word_start, 3),
+                        "text": "",
+                        "speaker": "PAUSE",
+                        "words": [],
+                        "pauses": [
+                            {
+                                "start": round(last_word_end, 3),
+                                "end": round(next_word_start, 3),
+                                "duration": round(next_word_start - last_word_end, 3)
+                            }
+                        ]
+                    }
+                    new_segments.append(pause_segment)
+    
+    data["segments"] = new_segments
+    
     with open(json_file, "w", encoding="utf-8") as f:
         json.dump(data, f, ensure_ascii=False, indent=4)
     
     print(f"Session {session_id} pause annotation done: {json_file}")
+    print(f"Total segments after processing: {len(new_segments)}")
+    
+    pause_segments = [seg for seg in new_segments if seg.get("speaker") == "PAUSE"]
+    print(f"Added {len(pause_segments)} inter-sentence pause segments")
+    
     return data
 
 if __name__ == "__main__":
-    annotated_data = annotate_pauses("000030", 0.1)
+    annotated_data = annotate_pauses("000030", 0.1)