Update app.py

app.py

@@ -23,73 +23,13 @@ import ffmpeg
 
 block = gr.Blocks()
 
-def pad_image(image):
-    w, h = image.size
-    if w == h:
-        return image
-    elif w > h:
-        new_image = Image.new(image.mode, (w, w), (0, 0, 0))
-        new_image.paste(image, (0, (w - h) // 2))
-        return new_image
-    else:
-        new_image = Image.new(image.mode, (h, h), (0, 0, 0))
-        new_image.paste(image, ((h - w) // 2, 0))
-        return new_image
-
-def calculate(image_in, audio_in):
-    waveform, sample_rate = torchaudio.load(audio_in)
-    waveform = torch.mean(waveform, dim=0, keepdim=True)
-    torchaudio.save("/content/audio.wav", waveform, sample_rate, encoding="PCM_S", bits_per_sample=16)
-    image = Image.open(image_in)
-    image = pad_image(image)
-    image.save("image.png")
-    print("Inside calculate")
-    return audio_in
-    pocketsphinx_run = subprocess.run(['pocketsphinx', '-phone_align', 'yes', 'single', '/content/audio.wav'], check=True, capture_output=True)
-    jq_run = subprocess.run(['jq', '[.w[]|{word: (.t | ascii_upcase | sub("<S>"; "sil") | sub("<SIL>"; "sil") | sub("\\\(2\\\)"; "") | sub("\\\(3\\\)"; "") | sub("\\\(4\\\)"; "") | sub("\\\[SPEECH\\\]"; "SIL") | sub("\\\[NOISE\\\]"; "SIL")), phones: [.w[]|{ph: .t | sub("\\\+SPN\\\+"; "SIL") | sub("\\\+NSN\\\+"; "SIL"), bg: (.b*100)|floor, ed: (.b*100+.d*100)|floor}]}]'], input=pocketsphinx_run.stdout, capture_output=True)
-    with open("test.json", "w") as f:
-        f.write(jq_run.stdout.decode('utf-8').strip())
-    
-    os.system(f"cd /content/one-shot-talking-face && python3 -B test_script.py --img_path /content/image.png --audio_path /content/audio.wav --phoneme_path /content/test.json --save_dir /content/train")
-    return "/content/train/image_audio.mp4"
+def cal(gender,input_text):
+    return gender+input_text
     
 def one_shot(image,input_text,gender): 
    if gender == 'Female' or gender == 'female':
-     print(gender,input_text)
-     tts = gTTS(input_text)
-     with tempfile.NamedTemporaryFile(suffix='.mp3', delete=False) as f:
-          tts.write_to_fp(f)
-          f.seek(0)
-          sound = AudioSegment.from_file(f.name, format="mp3")
-          sound.export("/content/audio.wav", format="wav")
-          audio_in="/content/audio.wav"
-     calculate(image,audio_in)
+     cal(gender,input_text)
      
-     
-       
-   elif gender == 'Male' or gender == 'male':
-      print(gender)
-      models, cfg, task = load_model_ensemble_and_task_from_hf_hub(
-          "Voicemod/fastspeech2-en-male1",
-          arg_overrides={"vocoder": "hifigan", "fp16": False}
-      )
-
-      model = models[0].cuda()
-      TTSHubInterface.update_cfg_with_data_cfg(cfg, task.data_cfg)
-      generator = task.build_generator([model], cfg)
-
-      sample = TTSHubInterface.get_model_input(task, input_text)
-      sample["net_input"]["src_tokens"] = sample["net_input"]["src_tokens"].cuda()
-      sample["net_input"]["src_lengths"] = sample["net_input"]["src_lengths"].cuda()
-      sample["speaker"] = sample["speaker"].cuda()
-
-      wav, rate = TTSHubInterface.get_prediction(task, model, generator, sample)
-      # soundfile.write("/content/audio_before.wav", wav, rate)
-      soundfile.write("/content/audio_before.wav", wav.cpu().clone().numpy(), rate)
-      cmd='ffmpeg -i /content/audio_before.wav -filter:a "atempo=0.7" -vn /content/audio.wav'
-      os.system(cmd)
-      one_shot_talking(image,'audio.wav')
-
        
 
 
@@ -106,7 +46,7 @@ def run():
           # audio_in = gr.Audio(show_label=False, type='filepath')
           input_text=gr.Textbox(lines=3, value="Hello How are you?", label="Input Text")
           gender = gr.Radio(["Female","Male"],value="Female",label="Gender")
-          video_out = gr.Audio(label="output")
+          video_out = gr.Textbox(label="output")
           # video_out = gr.Video(show_label=False)
         with gr.Row().style(equal_height=True):
           btn = gr.Button("Generate")