Update app.py

app.py

@@ -4,6 +4,12 @@ import gradio as gr
 import paddlehub as hub
 import urllib
 import cv2
+from datasets import load_dataset
+import re
+import os
+import requests
+
+from share_btn import community_icon_html, loading_icon_html, share_js
 
 import torch
 
@@ -95,37 +101,18 @@ def get_result(text_prompts, style_indx):
 
   interrogate_prompt = img_to_text(imagefile, fn_index=1)[0]
   print(interrogate_prompt)
-  music_output = get_music(interrogate_prompt + ", " + style_list_EN[style_indx])
+  spec_image, music_output = get_music(interrogate_prompt + ", " + style_list_EN[style_indx])
   
   video_merged = merge_video(music_output, image_output)
-  return {video_result:video_merged, status_text:'Success'}
+  return {spec_result:spec_image, video_result:video_merged, status_text:'Success'}
 
 def get_music(prompt):
-  
-  result = text_to_music(prompt, fn_index=0)
-  
-  print(f"""—————
-  NEW RESULTS
-  prompt : {prompt}
-  music : {result}
-  ———————
-  """)
-  
-  url = result
-  save_as = "file.mp3"
-  
-  data = urllib.request.urlopen(url)
-
-  f = open(save_as,'wb')
-  f.write(data.read())
-  f.close()
-  
-  #wave_file="file.wav"
-  
-  #sound = AudioSegment.from_mp3(save_as)
-  #sound.export(wave_file, format="wav")
-  
-  return save_as 
+  spec = pipe(prompt).images[0]
+  print(spec)
+  wav = wav_bytes_from_spectrogram_image(spec)
+  with open("output.wav", "wb") as f:
+    f.write(wav[0].getbuffer())
+  return spec, 'output.wav'
   
 def merge_video(music, img_list):
   #Convert to mp3   
@@ -446,6 +433,7 @@ with block:
             interactive=False
         )
         
+        spec_result = gr.Image()
         video_result = gr.Video(type=None, label='Final Merged video')
         
         trigger_component = gr.Textbox(vaule="", visible=False) # This component is used for triggering inference funtion.
@@ -458,14 +446,9 @@ with block:
         
         text.submit(translate_language, inputs=[text], outputs=[language_tips_text, status_text, trigger_component, translated_language])
         btn.click(translate_language, inputs=[text], outputs=[language_tips_text, status_text, trigger_component, translated_language])
-        trigger_component.change(fn=get_result, inputs=[translated_language, styles], outputs=[video_result, status_text])
+        trigger_component.change(fn=get_result, inputs=[translated_language, styles], outputs=[spec_result, video_result, status_text])
         
         
-        gr.Markdown(
-            """  
-### <u>[Prompt Tutorial 公式教程...](https://github.com/PaddlePaddle/PaddleHub/blob/develop/modules/image/text_to_image/ernie_vilg/README.md#四-prompt-指南)([Explore more...](https://github.com/PaddlePaddle/PaddleHub/blob/develop/modules/image/text_to_image/ernie_vilg/README.md#四-prompt-指南))</u>
-            """
-        )
         gr.Markdown(
             """  
       Space by [@DGSpitzer](https://www.youtube.com/channel/UCzzsYBF4qwtMwJaPJZ5SuPg)❤️ [@大谷的游戏创作小屋](https://space.bilibili.com/176003)