Update transcription_diarization.py

transcription_diarization.py

@@ -15,22 +15,30 @@ def upload_to_s3(local_file_path, bucket_name, s3_file_key):
     s3_client.upload_file(local_file_path, bucket_name, s3_file_key)
     return f's3://{bucket_name}/{s3_file_key}'
 
-def transcribe_video(file_uri, job_name):
+def transcribe_video(file_uri, job_name, language):
     transcribe = boto3.client('transcribe',
                               aws_access_key_id=aws_access_key_id,
                               aws_secret_access_key=aws_secret_access_key,
                               region_name='eu-central-1')
 
-    transcribe.start_transcription_job(
-        TranscriptionJobName=job_name,
-        Media={'MediaFileUri': file_uri},
-        MediaFormat='mp4',
-        IdentifyLanguage=True,
-        Settings={
+    language_code = get_language_code(language)
+
+    job_params = {
+        'TranscriptionJobName': job_name,
+        'Media': {'MediaFileUri': file_uri},
+        'MediaFormat': 'mp4',
+        'Settings': {
             'ShowSpeakerLabels': True,
             'MaxSpeakerLabels': 4
         }
-    )
+    }
+
+    if language_code:
+        job_params['LanguageCode'] = language_code
+    else:
+        job_params['IdentifyLanguage'] = True
+
+    transcribe.start_transcription_job(**job_params)
 
     while True:
         status = transcribe.get_transcription_job(TranscriptionJobName=job_name)
@@ -100,14 +108,27 @@ def extract_transcriptions_with_speakers(transcript_data):
 
     return transcriptions
 
-
-def diarize_audio(video_path):
+def get_language_code(language):
+    language_codes = {
+        "English": "en-US",
+        "Hebrew": "he-IL",
+        "Arabic": "ar-SA",
+        "French": "fr-FR",
+        "German": "de-DE",
+        "Italian": "it-IT",
+        "Japanese": "ja-JP",
+        "Chinese": "zh-CN",
+        "Auto-detect": None
+    }
+    return language_codes.get(language, "en-US") 
+
+def diarize_audio(video_path, language):
     bucket_name = 'transcriptionjobbucket'
     s3_file_key = os.path.basename(video_path)
     file_uri = upload_to_s3(video_path, bucket_name, s3_file_key)
 
     job_name = f'transcription_job_{int(time.time())}'
-    transcript_url = transcribe_video(file_uri, job_name)
+    transcript_url = transcribe_video(file_uri, job_name, language)
 
     if transcript_url:
         transcript_data = download_transcript(transcript_url)