Update app.py

app.py

@@ -24,7 +24,7 @@ def get_groq_response(prompt):
     client = Groq(api_key=os.environ["GROQ_API_KEY"])
     chat_completion = client.chat.completions.create(
         messages=[{"role": "user", "content": prompt}],
-        model="llama-3.3-70b-versatile"
+        model="llama3-70b-8192"
     )
     return chat_completion.choices[0].message.content
 
@@ -48,26 +48,6 @@ def generate_script(topic, duration):
       return ""
 
 
-def translate_to_urdu(english_script):
-    try:
-        tokenizer.src_lang = "en"
-        inputs = tokenizer(english_script, return_tensors="pt", truncation=True, max_length=1024).to(model.device)
-
-        translated_tokens = model.generate(
-            **inputs,
-            max_length=1024,
-            no_repeat_ngram_size=2,  # Reduce repetition checks (faster)
-            forced_bos_token_id=tokenizer.get_lang_id("ur"),
-            num_beams=2  # Faster than beam search with high values
-        )
-
-        urdu_script = tokenizer.decode(translated_tokens[0], skip_special_tokens=True)
-        return urdu_script
-
-    except Exception as e:
-        return f"❌ Error in translation: {str(e)}"
-
-# # ✅ Function to Translate English Script to Urdu
 # def translate_to_urdu(english_script):
 #     try:
 #         tokenizer.src_lang = "en"
@@ -77,38 +57,70 @@ def translate_to_urdu(english_script):
 #             **inputs,
 #             max_length=1024,
 #             no_repeat_ngram_size=2,  # Reduce repetition checks (faster)
-#             forced_bos_token_id=tokenizer.lang_code_to_id["ur"],  # Fix for language ID error
-#             num_beams=2  # Faster translation
+#             forced_bos_token_id=tokenizer.get_lang_id("ur"),
+#             num_beams=2  # Faster than beam search with high values
 #         )
 
 #         urdu_script = tokenizer.decode(translated_tokens[0], skip_special_tokens=True)
 #         return urdu_script
+
 #     except Exception as e:
 #         return f"❌ Error in translation: {str(e)}"
 
+def translate_to_urdu(english_script):
+    try:
+        tokenizer.src_lang = "en"
+        max_length = 500  # Process smaller chunks to avoid truncation
+        input_chunks = [english_script[i:i+max_length] for i in range(0, len(english_script), max_length)]
+
+        translated_chunks = []
+        for chunk in input_chunks:
+            inputs = tokenizer(chunk, return_tensors="pt", truncation=True, max_length=1024).to(model.device)
+            translated_tokens = model.generate(
+                **inputs,
+                max_length=1024,
+                no_repeat_ngram_size=2,
+                forced_bos_token_id=tokenizer.get_lang_id("ur"),
+                num_beams=2
+            )
+            translated_text = tokenizer.batch_decode(translated_tokens, skip_special_tokens=True)[0]
+            translated_chunks.append(translated_text)
+
+        return " ".join(translated_chunks)
+
+    except Exception as e:
+        return f"❌ Error in translation: {str(e)}"
+
 
 
-# ✅ Function to Save Edited Urdu Script to a File
 def save_edited_urdu(edited_urdu_script, topic):
     timestamp = datetime.datetime.now().strftime("%Y_%m_%d")
     filename = f"{topic}_Urdu_Final_{timestamp}.txt"
+    filepath = os.path.join(os.getcwd(), filename)  # Ensure full path
 
-    with open(filename, "w", encoding="utf-8") as file:
+    with open(filepath, "w", encoding="utf-8") as file:
         file.write(edited_urdu_script)
 
-    return filename  # Return the file name so it can be downloaded
+    return filepath  # Return full path for Gradio download
 
 
 
 
 def save_file(content, filename):
-    with open(filename, "w", encoding="utf-8") as file:
+    timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")  # Add seconds to avoid overwriting
+    filename = f"{filename}_{timestamp}.txt"  # Ensure uniqueness
+    filepath = os.path.join(os.getcwd(), filename)
+
+    with open(filepath, "w", encoding="utf-8") as file:
         file.write(content)
-    return filename
+
+    return filepath
+
 
 def process_request(topic, duration):
     eng_script = generate_script(topic, duration)
-    return eng_script, ""
+    return eng_script, "", ""  # Clears Urdu and Edited Urdu textboxes
+
 
 def process_translation(eng_script):
     urdu_script = translate_to_urdu(eng_script)
@@ -137,10 +149,13 @@ with gr.Blocks() as app:
     duration_input = gr.Slider(minimum=1, maximum=30, step=1, label="Duration (minutes)")
     generate_button = gr.Button("Generate English Script")
     eng_output = gr.Textbox(label="Generated English Script", interactive=False)
-    generate_button.click(process_request, inputs=[topic_input, duration_input], outputs=[eng_output])
-
+    generate_button.click(
+        process_request,
+        inputs=[topic_input, duration_input],
+        outputs=[eng_output, urdu_output, editable_urdu]  # Clears Urdu & Edited Urdu textboxes
+    )
     translate_button = gr.Button("Generate Urdu Script")
-    urdu_output = gr.Textbox(label="Translated Urdu Script (Editable)", interactive=True)
+    urdu_output = gr.Textbox(label="Translated Urdu Script", interactive=True)
     translate_button.click(process_translation, inputs=[eng_output], outputs=[urdu_output])
 
     # Editable Urdu Textbox
@@ -169,6 +184,10 @@ with gr.Blocks() as app:
 
     download_eng_button.click(download_english, inputs=[eng_output, topic_input], outputs=[download_eng_file])
     download_urdu_button.click(download_urdu, inputs=[urdu_output, topic_input], outputs=[download_urdu_file])
+    # Save the edited script
+    save_button.click(save_edited_urdu, inputs=[editable_urdu, topic_input], outputs=[download_final_urdu_file])
+
+
     download_final_urdu_button.click(download_final_urdu, inputs=[editable_urdu, topic_input], outputs=[download_final_urdu_file])
 
-app.launch()
+app.launch()