Update app.py

app.py

@@ -37,39 +37,90 @@
 # iface.launch()
 
 
+# import gradio as gr
+# import requests
+# import urllib.request
+# from pydub import AudioSegment
+# import numpy as np
+# import os
+
+# def function1(prompt):
+#     response = requests.post("https://tommy24-testing3.hf.space/run/predict", json={
+#       "data": [
+#         prompt,
+#     ]}).json()
+#     data = response["data"][0]
+#     response = requests.post("https://matthijs-speecht5-tts-demo.hf.space/run/predict", json={
+#         "data": [
+#             data,
+#             "KSP (male)",
+#         ]
+#     }).json()
+#     data = response["data"][0]["name"]
+#     data = "https://matthijs-speecht5-tts-demo.hf.space/file="+data
+#     file_name, headers = urllib.request.urlretrieve(data, "speech.mp3")
+#     # code = random.randint(1,1000)
+#     # generated_file = f"output{code}"
+#     filename = "output.mp3"
+    
+#     if os.path.exists(filename):
+#         os.remove(filename)
+#     else:
+#         pass
+#     command = f"ffmpeg -i {file_name} -vn -ar 44100 -ac 2 -b:a 192k output.mp3"
+#     os.system(command)
+#     return "output.mp3"
+
+# iface = gr.Interface(fn=function1, inputs="text", outputs=[gr.Audio(label="Audio",type="numpy")])
+# iface.launch()
+
 import gradio as gr
 import requests
 import urllib.request
 from pydub import AudioSegment
 import numpy as np
 import os
+import sys
+import wave
+import io
+import base64
+import azure.cognitiveservices.speech as speechsdk
+
+speech_key = os.environ.get("test3")
+service_region = os.environ.get("test4")
+
+speech_config = speechsdk.SpeechConfig(subscription=speech_key, region=service_region)
+# Note: the voice setting will not overwrite the voice element in input SSML.
+speech_config.speech_synthesis_voice_name = os.environ.get("test5")
 
 def function1(prompt):
     response = requests.post("https://tommy24-testing3.hf.space/run/predict", json={
       "data": [
         prompt,
     ]}).json()
-    data = response["data"][0]
-    response = requests.post("https://matthijs-speecht5-tts-demo.hf.space/run/predict", json={
-        "data": [
-            data,
-            "KSP (male)",
-        ]
-    }).json()
-    data = response["data"][0]["name"]
-    data = "https://matthijs-speecht5-tts-demo.hf.space/file="+data
-    file_name, headers = urllib.request.urlretrieve(data, "speech.mp3")
-    # code = random.randint(1,1000)
-    # generated_file = f"output{code}"
-    filename = "output.mp3"
+    message = response["data"][0]
+    speech_synthesizer = speechsdk.SpeechSynthesizer(speech_config=speech_config)
+    result = speech_synthesizer.speak_text_async(message).get()
+    if result.reason == speechsdk.ResultReason.SynthesizingAudioCompleted:
+        audio_stream = io.BytesIO(result.audio_data)
     
-    if os.path.exists(filename):
-        os.remove(filename)
-    else:
-        pass
-    command = f"ffmpeg -i {file_name} -vn -ar 44100 -ac 2 -b:a 192k output.mp3"
-    os.system(command)
-    return "output.mp3"
+        # Create a wave file object and write the audio data to it
+        with wave.open("audio.wav", 'wb') as wave_file:
+            wave_file.setnchannels(1)
+            wave_file.setsampwidth(2)
+            wave_file.setframerate(16000)
+            wave_file.writeframesraw(audio_stream.getvalue())
+        
+        # Use ffmpeg to convert the wave file to an mp3 file
+        filename = "output.mp3"
+        
+        if os.path.exists(filename):
+            os.remove(filename)
+        else:
+            pass
+        command = f"ffmpeg -i audio.wav -y -codec:a libmp3lame -qscale:a 2 {filename}"
+        os.system(command)
+        return "output.mp3"
 
 iface = gr.Interface(fn=function1, inputs="text", outputs=[gr.Audio(label="Audio",type="numpy")])
 iface.launch()