Update app.py

app.py

@@ -2,9 +2,15 @@ import gradio as gr
 import os 
 import json 
 import requests
-from gradio.inputs import SpeechToText
+from google.cloud import speech_v1p1beta1 as speech
+from google.oauth2 import service_account
+import base64
+import io
 
+path_to_key_file = "/path/to/your/key-file.json"
 
+credentials = service_account.Credentials.from_service_account_file(path_to_key_file)
+client = speech.SpeechClient(credentials=credentials)
 
 #Streaming endpoint 
 API_URL = "https://api.openai.com/v1/chat/completions" #os.getenv("API_URL") + "/generate_stream"
@@ -12,6 +18,22 @@ API_URL = "https://api.openai.com/v1/chat/completions" #os.getenv("API_URL") + "
 #Huggingface provided GPT4 OpenAI API Key 
 OPENAI_API_KEY = os.getenv("OPENAI_API_KEY") 
 
+def transcribe_audio(audio):
+    audio_bytes = base64.b64decode(audio.split(",")[1])
+    audio_file = io.BytesIO(audio_bytes)
+
+    audio = speech.RecognitionAudio(content=audio_file.read())
+    config = speech.RecognitionConfig(
+        encoding=speech.RecognitionConfig.AudioEncoding.LINEAR16,
+        sample_rate_hertz=16000,
+        language_code="en-US",
+    )
+
+    response = client.recognize(config=config, audio=audio)
+
+    for result in response.results:
+        return result.alternatives[0].transcript
+
 #Inferenec function
 def predict(system_msg, inputs, top_p, temperature, chat_counter, chatbot=[], history=[]):  
 
@@ -129,7 +151,7 @@ with gr.Blocks(css = """#col_container { margin-left: auto; margin-right: auto;}
             system_msg = gr.Textbox(label="Instruct the AI Assistant to set its beaviour", info = system_msg_info, value="")
             accordion_msg = gr.HTML(value="🚧 To set System message you will have to refresh the app", visible=False)
         chatbot = gr.Chatbot(label='GPT4', elem_id="chatbot")
-        inputs = SpeechToText(placeholder="Hi there!", label="Record an input or type a text")
+        inputs = gr.Audio(label="Record an input", source="microphone")
         state = gr.State([]) 
         with gr.Row():
             with gr.Column(scale=7):
@@ -144,8 +166,9 @@ with gr.Blocks(css = """#col_container { margin-left: auto; margin-right: auto;}
             chat_counter = gr.Number(value=0, visible=False, precision=0)
 
     #Event handling
-    inputs.submit( predict, [system_msg, inputs, top_p, temperature, chat_counter, chatbot, state], [chatbot, state, chat_counter, server_status_code],)  #openai_api_key
-    b1.click( predict, [system_msg, inputs, top_p, temperature, chat_counter, chatbot, state], [chatbot, state, chat_counter, server_status_code],)  #openai_api_key
+    inputs.submit(predict, [system_msg, transcribe_audio(inputs), top_p, temperature, chat_counter, chatbot, state], [chatbot, state, chat_counter, server_status_code])
+    b1.click(predict, [system_msg, transcribe_audio(inputs), top_p, temperature, chat_counter, chatbot, state], [chatbot, state, chat_counter, server_status_code])
+
     
     inputs.submit(set_visible_false, [], [system_msg])
     b1.click(set_visible_false, [], [system_msg])