Update app.py

app.py

@@ -180,21 +180,24 @@ def transcribe_and_diarize(file_path, task, progress=gr.Progress()):
 def format_to_markdown(transcription_text, speaker_transcription, audio_duration=None, location=None, speaker_age=None, context=None, num_speakers=None, speaker_names=None):
     metadata = {
         "Date de traitement": datetime.now().strftime('%d/%m/%Y %H:%M'),
-        "Durée de l'audio": f"{audio_duration} secondes" if audio_duration else "[à remplir]",
+        "Durée de l'audio": f"{audio_duration}" if audio_duration else "[à remplir]",
         "Lieu": location if location else "[non spécifié]",
         "Nombre d'interlocuteurs": num_speakers if num_speakers else "[non spécifié]",
         "Âge de l'intervenant": f"{speaker_age} ans" if speaker_age else "[non spécifié]",
         "Contexte": context if context else "[non spécifié]"
     }
     
-    # Gestion plus robuste de speaker_names
+    # Créer un dictionnaire pour les noms personnalisés
+    custom_names = {}
     if speaker_names is not None:
-        if isinstance(speaker_names, list):
-            for item in speaker_names:
-                if isinstance(item, (list, tuple)) and len(item) == 2:
-                    metadata[item[0]] = item[1]
-                elif isinstance(item, dict) and "Speaker ID" in item and "Nom personnalisé" in item:
-                    metadata[item["Speaker ID"]] = item["Nom personnalisé"]
+        for row in speaker_names:
+            if isinstance(row, dict):
+                speaker_id = row.get("Speaker ID")
+                custom_name = row.get("Nom personnalisé")
+                if speaker_id and custom_name:
+                    custom_names[speaker_id] = custom_name
+            elif isinstance(row, (list, tuple)) and len(row) == 2:
+                custom_names[row[0]] = row[1]
     
     metadata_text = "\n".join([f"- **{key}** : '{value}'" for key, value in metadata.items()])
     
@@ -202,14 +205,11 @@ def format_to_markdown(transcription_text, speaker_transcription, audio_duration
         if isinstance(speaker_transcription, str):
             speaker_transcription = parse_simplified_diarization(speaker_transcription)
         
-        if isinstance(speaker_transcription, list) and all(isinstance(item, tuple) and len(item) == 2 for item in speaker_transcription):
-            formatted_transcription = []
-            for speaker, text in speaker_transcription:
-                custom_name = next((item[1] for item in speaker_names if item[0] == speaker), speaker) if isinstance(speaker_names, list) else speaker
-                formatted_transcription.append(f"**{custom_name}**: {text}")
-            transcription_text = "\n\n".join(formatted_transcription)
-        else:
-            raise ValueError("Invalid speaker transcription format")
+        formatted_transcription = []
+        for speaker, text in speaker_transcription:
+            custom_name = custom_names.get(speaker, speaker)
+            formatted_transcription.append(f"**{custom_name}**: {text}")
+        transcription_text = "\n\n".join(formatted_transcription)
     except Exception as e:
         print(f"Error formatting speaker transcription: {e}")
         transcription_text = "Error formatting speaker transcription. Using raw transcription instead.\n\n" + transcription_text