Upload folder using huggingface_hub

app.py

@@ -18,7 +18,7 @@ Note:
 import gradio as gr
 import tempfile
 import openai
-from typing import Tuple
+from typing import List
 
 
 def tts(input_text: str, model: str, voice: str, api_key: str) -> str:
@@ -47,49 +47,57 @@ def tts(input_text: str, model: str, voice: str, api_key: str) -> str:
     openai.api_key = api_key
 
     try:
-        response = openai.audio.speech.create(
-            input=input_text,
+        response = openai.Audio.create(
+            text=input_text,
             voice=voice,
             model=model
         )
     except openai.error.OpenAIError as e:
         raise e
 
-    if not hasattr(response, 'content'):
+    if not hasattr(response, 'audio'):
         raise Exception("Invalid response from OpenAI API. The response does not contain audio content.")
 
     with tempfile.NamedTemporaryFile(suffix=".mp3", delete=False) as temp_file:
-        temp_file.write(response.content)
+        temp_file.write(response.audio)
         temp_file_path = temp_file.name
 
     return temp_file_path
 
 
-def on_convert_click(input_text: str, model: str, voice: str, api_key: str) -> Tuple[str, str]:
+def tts_batch(
+    input_texts: List[str],
+    model: str,
+    voice: str,
+    api_key: str,
+) -> List[str]:
     """
-    Callback function to handle the click event for text-to-speech conversion.
+    Convert a batch of input texts to speech using OpenAI's Text-to-Speech API.
 
-    :param input_text: Text input from the user.
-    :type input_text: str
-    :param model: Selected model.
+    :param input_texts: The texts to be converted to speech.
+    :type input_texts: List[str]
+    :param model: The model to use for synthesis.
     :type model: str
-    :param voice: Selected voice.
+    :param voice: The voice profile to use.
     :type voice: str
-    :param api_key: User's OpenAI API key.
+    :param api_key: OpenAI API key.
     :type api_key: str
-    :return: Tuple containing the file path to the generated audio file and an error message.
-    :rtype: Tuple[str, str]
+    :return: List of file paths to the generated audio files.
+    :rtype: List[str]
     """
-    try:
-        file_path = tts(input_text, model, voice, api_key)
-        return file_path, ""
-    except Exception as e:
-        return None, str(e)
+    outputs = []
+    for input_text in input_texts:
+        try:
+            output = tts(input_text, model, voice, api_key)
+            outputs.append(output)
+        except Exception as e:
+            outputs.append(None)
+    return outputs
 
 
 def main():
     """
-    Main function to create and launch the Gradio interface.
+    Main function to create and launch the Gradio interface with API capabilities and enhancements.
     """
     # Define model and voice options
     MODEL_OPTIONS = ["tts-1", "tts-1-hd"]
@@ -120,22 +128,42 @@ def main():
                     label="Error Message", interactive=False, visible=False
                 )
 
-        # Define the event handler for the submit button
+        # Define the event handler for the submit button with API endpoint and naming
         submit_button.click(
-            fn=on_convert_click,
+            fn=tts,
             inputs=[input_textbox, model_dropdown, voice_dropdown, api_key_input],
-            outputs=[output_audio, error_output]
+            outputs=output_audio,
+            api_name="convert_text_to_speech",
         )
 
         # Allow pressing Enter in the input textbox to trigger the conversion
         input_textbox.submit(
-            fn=on_convert_click,
+            fn=tts,
             inputs=[input_textbox, model_dropdown, voice_dropdown, api_key_input],
-            outputs=[output_audio, error_output]
+            outputs=output_audio,
+            api_name="convert_text_to_speech",
+        )
+
+    # Expose the `demo` app as a callable function
+    def process_text_to_speech(
+        input_text: str,
+        model: str = "tts-1",
+        voice: str = "echo",
+        api_key: str = ""
+    ) -> str:
+        """
+        Allows calling the Gradio app as a function.
+        """
+        return demo(
+            input_text,
+            model,
+            voice,
+            api_key
         )
 
-    demo.launch()
+    # Launch the Gradio app with API documentation enabled
+    demo.launch(share=True)
 
 
 if __name__ == "__main__":
-    main()
+    main()