add API

app.py

@@ -11,6 +11,7 @@ import torch
 import torchaudio
 from transformers import pipeline
 from pathlib import Path
+from flask import Flask, send_file
 
 # local import
 import sys
@@ -151,10 +152,6 @@ def ASRTTS_clean(audio_file, spk_name):
     return save_id
 
 
-# def ref_reg_callback(audio_file, spk_name, ref_text):
-#     reg_text = ref_text
-#     return audio_file, spk_name, reg_text
-
 reference_textbox = gr.Textbox(
     value="",
     placeholder="Input reference here",
@@ -168,17 +165,6 @@ recognization_textbox = gr.Textbox(
 )
 
 speaker_option = gr.Radio(choices=spk_names, label="Speaker")
-# speaker_profiles = {
-#     "Male_1": "speaker_icons/male1.png",
-#     "Male_2": "speaker_icons/male2.png",
-#     "Female_1": "speaker_icons/female1.png",
-#     "Female_2": "speaker_icons/female2.png",
-# }
-
-# speaker_option = gr.Image(label="Choose your speaker profile",
-#                                  image_mode="RGB",
-#                                  options=speaker_profiles
-#                                  )
 
 input_audio = gr.Audio(
     source="upload", type="filepath", label="Audio_to_Evaluate"
@@ -215,21 +201,35 @@ def show_icon(choice):
         spk_icon = gr.Image.update(value="speaker_icons/female2.png", visible=True)
     return spk_icon
 
+def get_download_file(audio_file=None):
+    if audio_file == None:
+        output_audio_file = gr.File.update(visible=False)
+    else:
+        output_audio_file = gr.File.update(visible=True)
+    return output_audio_file
+        
+def download_file(audio_file):
+    return gr.File(value=audio_file)
+# pdb.set_trace()
+
+# if __name__ == "__main__":
+#     file_share_app.run(port=3000)    
 
 with gr.Blocks(
     analytics_enabled=False,
     css=".gradio-container {background-color: #78BD91}",
 ) as demo:
-    with gr.Column():
+    with gr.Column(elem_id="Column"):
         input_format = gr.Radio(
-            choices=["microphone", "upload"], label="Choose your input format"
+            choices=["microphone", "upload"], label="Choose your input format", elem_id="input_format"
         )
         input_audio = gr.Audio(
-            source="upload",
+            source="microphone",
             type="filepath",
             label="Input Audio",
             interactive=True,
             visible=False,
+            elem_id="input_audio"
         )
         input_format.change(
             fn=change_audiobox, inputs=input_format, outputs=input_audio
@@ -248,14 +248,18 @@ with gr.Blocks(
         )
 
     b2 = gr.Button("Convert")
-
+    
     output_audio = gr.Audio(
-        source="upload", file="filepath", label="Converted Audio"
+        source="upload", file="filepath", label="Converted Audio", interactive=False
     )
+    
     b2.click(
         ASRTTS_clean,
         inputs=[input_audio, speaker_option],
         outputs=output_audio,
+        api_name="convert"
     )
+    
+# download_file("wav/001_F1_spkembs.wav")
 
 demo.launch(share=False)