Update app.py

app.py

@@ -153,30 +153,36 @@ def get_translation_model(source_language, target_language):
     # Return the model using string concatenation
     return f"Helsinki-NLP/opus-mt-{source_language}-{target_language}"
 
+def translate_single_entry(entry, translator, target_language):
+    original_text = entry["text"]
+    translated_text = translator(original_text, target_language=target_language)[0]['translation_text']
+    return {
+        "start": entry["start"],
+        "original": original_text,
+        "translated": translated_text,
+        "end": entry["end"]
+    }
+
 def translate_text(transcription_json, source_language, target_language):
     # Load the translation model for the specified target language
     translation_model_id = get_translation_model(source_language, target_language)
     logger.debug(f"Translation model: {translation_model_id}")
     translator = pipeline("translation", model=translation_model_id)
 
-    # Prepare output structure
-    translated_json = []
-
-    # Translate each sentence and store it with its start time
-    for entry in transcription_json:
-        original_text = entry["text"]
-        translated_text = translator(original_text)[0]['translation_text']
-        translated_json.append({
-            "start": entry["start"],
-            "original": original_text,
-            "translated": translated_text,
-            "end": entry["end"]
-        })
-        # Log the components being added to translated_json
-        logger.debug("Adding to translated_json: start=%s, original=%s, translated=%s, end=%s",
-                     entry["start"], original_text, translated_text, entry["end"])
-
-    # Return the translated timestamps as a JSON string
+    # Use ThreadPoolExecutor to parallelize translations
+    with concurrent.futures.ThreadPoolExecutor() as executor:
+        # Submit all translation tasks and collect results
+        translate_func = lambda entry: translate_single_entry(entry, translator, target_language)
+        translated_json = list(executor.map(translate_func, transcription_json))
+
+    # Sort the translated_json by start time
+    translated_json.sort(key=lambda x: x["start"])
+
+    # Log the components being added to translated_json
+    for entry in translated_json:
+        logger.debug("Added to translated_json: start=%s, original=%s, translated=%s, end=%s",
+                     entry["start"], entry["original"], entry["translated"], entry["end"])
+
     return translated_json
 
 def update_translations(file, edited_table, mode):