Update app.py

app.py

@@ -1,17 +1,45 @@
-import streamlit as st
-import speech_recognition as sr
 import os
-from groq import Groq
+import sounddevice as sd
+import queue
+import vosk
+import json
 from gtts import gTTS
 import tempfile
 import playsound
+from groq import Groq
+import streamlit as st
 
-# Initialize Groq client
+# Set Groq API key
 GROQ_API_KEY = "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"  # Replace with your Groq API key
 os.environ["GROQ_API_KEY"] = GROQ_API_KEY
 client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
 
-# Function to use Groq API for grammar correction
+# Initialize Vosk model
+vosk_model_path = "vosk-model-small-en-us-0.15"  # Replace with your downloaded Vosk model path
+if not os.path.exists(vosk_model_path):
+    raise FileNotFoundError(f"Vosk model not found at {vosk_model_path}. Download it first.")
+model = vosk.Model(vosk_model_path)
+q = queue.Queue()
+
+# Function to handle audio input
+def audio_callback(indata, frames, time, status):
+    if status:
+        print(status)
+    q.put(bytes(indata))
+
+# Function to capture speech
+def get_speech():
+    with sd.RawInputStream(samplerate=16000, blocksize=8000, dtype="int16",
+                           channels=1, callback=audio_callback):
+        recognizer = vosk.KaldiRecognizer(model, 16000)
+        st.write("Listening... Speak now!")
+        while True:
+            data = q.get()
+            if recognizer.AcceptWaveform(data):
+                result = json.loads(recognizer.Result())
+                return result.get("text", "")
+
+# Function to correct grammar using Groq
 def correct_grammar_with_groq(text):
     try:
         chat_completion = client.chat.completions.create(
@@ -30,45 +58,26 @@ def correct_grammar_with_groq(text):
 
 # Function to speak text using gTTS
 def speak_text(text):
-    try:
-        # Try using pyttsx3
-        import pyttsx3
-        engine = pyttsx3.init()
-        engine.say(text)
-        engine.runAndWait()
-    except OSError:
-        # Fallback to gTTS
-        tts = gTTS(text=text, lang='en')
-        with tempfile.NamedTemporaryFile(delete=True, suffix=".mp3") as fp:
-            tts.save(fp.name)
-            playsound.playsound(fp.name)
-
-# Function to process speech input
-def get_speech():
-    try:
-        recognizer = sr.Recognizer()
-        with sr.Microphone() as source:
-            st.write("Listening... Speak now!")
-            audio = recognizer.listen(source, timeout=5)
-            return recognizer.recognize_google(audio)
-    except (AttributeError, OSError):
-        st.warning("Microphone not found or PyAudio not installed. Please type your input instead.")
-        return st.text_input("Enter your sentence:")
+    tts = gTTS(text=text, lang='en')
+    with tempfile.NamedTemporaryFile(delete=True, suffix=".mp3") as fp:
+        tts.save(fp.name)
+        playsound.playsound(fp.name)
 
 # Streamlit interface
 st.title("Spoken English Practice")
-st.write("Speak into your mic, or type your sentence if the mic is unavailable.")
+st.write("Speak into your mic, and I'll help you improve your English.")
+
 # Button to start speech
 if st.button("Speak Now"):
     user_input = get_speech()
     if user_input:
         st.write(f"You said: {user_input}")
 
-        # Grammar correction using Groq API
-        corrected = correct_grammar_with_groq(user_input)
-        st.write(f"Corrected: {corrected}")
+        # Correct grammar using Groq API
+        corrected_text = correct_grammar_with_groq(user_input)
+        st.write(f"Corrected: {corrected_text}")
 
         # Speak corrected text
-        speak_text(corrected)
+        speak_text(corrected_text)
 
-st.write("Ready to practice? Click 'Speak Now'!")
+st.write("Ready to practice? Click 'Speak Now'!")