bug fixed now... not done...

pages/main.py

@@ -94,10 +94,9 @@ def read_file(file_name: str) -> str:
 
 
 
-def split_audio(audio_file: str, 
-                max_size_mb: int = 25) -> List[str]:
+def split_audio(audio_file: str, max_size_mb: int = 25) -> List[str]:
     """
-    Divise un fichier audio en segments de taille maximale spécifiée et retourne les chemins des segments.
+    Divise un fichier audio en segments de taille maximale spécifiée.
 
     Args:
         audio_file (str): Chemin vers le fichier audio.
@@ -109,10 +108,7 @@ def split_audio(audio_file: str,
     try:
         audio = AudioSegment.from_wav(audio_file)
         duration_ms = len(audio)
-        segment_duration_ms = int(
-            (max_size_mb * 1024 * 1024 * 8) /
-            (audio.frame_rate * audio.sample_width * audio.channels)
-        )
+        segment_duration_ms = int((max_size_mb * 1024 * 1024 * 8) / (audio.frame_rate * audio.sample_width * audio.channels))
 
         segments = []
         with tempfile.TemporaryDirectory() as temp_dir:
@@ -123,39 +119,34 @@ def split_audio(audio_file: str,
                 segment.export(temp_segment_path, format="wav")
                 segments.append(temp_segment_path)
 
-
-                # Start of Selection
-                # Traiter les segments ici
-                language = st.session_state.get('target_language', 'en')  # Définir la variable
-                transcriptions = []
-                for segment_path in segments:
-                    try:
-                        with open(segment_path, "rb") as audio_segment:
-                            transcript = client.audio.transcriptions.create(
-                                model="whisper-1",
-                                file=audio_segment,
-                                # Start of Selection
-                                language=language,
-                                response_format="text"
-                            )
-                            transcriptions.append(transcript.text)
-                    except Exception as e:
-                        print(f"Erreur lors de la transcription du segment {segment_path} : {e}")
-                    finally:
-                        os.remove(segment_path)  # Supprime le segment après traitement
-                # Fin du traitement des segments et suppression automatique du répertoire temporaire
-
         return segments
-    except IOError as e:
-        print(f"Erreur lors de la lecture ou de l'écriture du fichier audio : {e}")
-        return []
-    except ValueError as e:
-        print(f"Erreur de valeur lors du traitement de l'audio : {e}")
-        return []
     except Exception as e:
-        print(f"Une erreur inattendue s'est produite : {e}")
+        print(f"Erreur lors du découpage de l'audio : {e}")
         return []
 
+def transcribe_segment(segment_path: str, language: Optional[str] = None) -> str:
+    """
+    Transcrit un segment audio en texte.
+
+    Args:
+        segment_path (str): Chemin vers le segment audio.
+        language (Optional[str]): La langue de l'audio.
+
+    Returns:
+        str: Le texte transcrit.
+    """
+    try:
+        with open(segment_path, "rb") as audio_segment:
+            transcript = client.audio.transcriptions.create(
+                model="whisper-1",
+                file=audio_segment,
+                language=language, # semble que language soit mal formatter au format ISO6391
+                response_format="text"
+            )
+        return transcript
+    except Exception as e:
+        print(f"Erreur lors de la transcription du segment {segment_path} : {e}")
+        return ""
 
 def transcribe_audio(audio_file: Union[str, IO], language: Optional[str] = None) -> str:
     """
@@ -172,13 +163,8 @@ def transcribe_audio(audio_file: Union[str, IO], language: Optional[str] = None)
     
     try:
         with st.status("Transcription de l'audio en cours...") as status:
-            # Si audio_file est une chaîne, on l'ouvre comme un fichier
-            if isinstance(audio_file, str):
-                file_size_mb = os.path.getsize(audio_file) / (1024 * 1024)
-                audio_file_path = audio_file
-            else:
-                file_size_mb = os.path.getsize(audio_file.name) / (1024 * 1024)
-                audio_file_path = audio_file.name
+            audio_file_path = audio_file if isinstance(audio_file, str) else audio_file.name
+            file_size_mb = os.path.getsize(audio_file_path) / (1024 * 1024)
 
             if file_size_mb > max_size_mb:
                 status.update(label="Découpage de l'audio en segments...")
@@ -186,30 +172,15 @@ def transcribe_audio(audio_file: Union[str, IO], language: Optional[str] = None)
                 full_transcript = ""
                 for i, segment in enumerate(segments):
                     status.update(label=f"Transcription du segment {i+1}/{len(segments)}...")
-                    with open(segment, "rb") as audio_segment:
-                        transcript = client.audio.transcriptions.create(
-                            model="whisper-1",
-                            file=audio_segment,
-                            language=language,
-                            response_format="text"
-                        )
-                        full_transcript += f"{transcript} "
+                    transcript = transcribe_segment(segment, language)
+                    full_transcript += f"{transcript} "
                 status.update(label="Transcription terminée", state="complete")
                 return full_transcript.strip()
             else:
                 status.update(label="Transcription de l'audio...")
-                with open(audio_file_path, "rb") as audio_file:
-                    transcript = client.audio.transcriptions.create(
-                        model="whisper-1",
-                        file=audio_file,
-                        language=language,
-                        response_format="text"
-                    )
+                transcript = transcribe_segment(audio_file_path, language)
                 status.update(label="Transcription terminée", state="complete")
                 return transcript
-    except IOError as e:
-        st.error(f"Erreur d'entrée/sortie lors de la transcription : {e}")
-        return ""
     except Exception as e:
         st.error(f"Erreur lors de la transcription : {e}")
         return ""