app.py

import streamlit as st
import joblib as j
import re as rg
from nltk.stem import PorterStemmer
import speech_recognition as sr

import os
import gdown
import joblib as j
'''
model_file = 'random_forest_model.pkl'
model_url = 'https://drive.google.com/uc?id=1suMJ0qgG5-oLhm_Jmyy7sst_myuiI2zg'

if not os.path.exists(model_file):
    print("Model file not found. Downloading from Google Drive...")
    gdown.download(model_url, model_file, quiet=False)
else:
    print("Model file found. Proceeding to load.")

model = j.load(model_file)
'''
# -----------------------------------------------------------
# 1) Load Vectorizer & Model
# -----------------------------------------------------------
vectorizer = j.load('tfidf_vectorizer.pkl')
model = j.load('random_forest_model.pkl')

pt = PorterStemmer()

# -----------------------------------------------------------
# 2) Preprocessing Function
# -----------------------------------------------------------
def preprocessing(text):
    text = rg.sub('[^a-zA-Z0-9\\s]', '', text.lower())
    words = [pt.stem(word) for word in text.split()]
    return " ".join(words)

# -----------------------------------------------------------
# 3) Prediction Function
# -----------------------------------------------------------
def predict(text):
    preprocessed_text = preprocessing(text)
    vector = vectorizer.transform([preprocessed_text])
    return model.predict(vector)[0]

# -----------------------------------------------------------
# 4) Speech Recognition
# -----------------------------------------------------------
def recognize_speech():
    recognizer = sr.Recognizer()
    with sr.Microphone() as source:
        st.write("Listening... Please speak now.")
        try:
            audio = recognizer.listen(source, timeout=5)
            text = recognizer.recognize_google(audio)
            return text
        except sr.WaitTimeoutError:
            st.write("Listening timed out while waiting for phrase to start.")
        except sr.UnknownValueError:
            st.write("Sorry, could not understand the audio.")
        except sr.RequestError as e:
            st.write(f"Could not request results from Google Speech Recognition service; {e}")
        return ""

# -----------------------------------------------------------
# 5) Set up Session State
# -----------------------------------------------------------
# Initialize a key in session_state to store spoken text
if "spoken_text" not in st.session_state:
    st.session_state["spoken_text"] = ""

# -----------------------------------------------------------
# 6) Streamlit UI
# -----------------------------------------------------------
st.title("Mental Health Sentiment Analysis")

# Radio for input method
input_option = st.radio("Choose input method:", ("Type Text", "Speak Text"))

if input_option == "Type Text":
    # Use a local variable for typed text
    input_text = st.text_area("Enter text for sentiment analysis:")
    
    if st.button("Predict"):
        if input_text.strip():
            result = predict(input_text)
            st.write(f"Predicted Sentiment: {result}")
        else:
            st.write("Please enter some text.")

elif input_option == "Speak Text":
    # Display what is currently stored in session_state
    if st.session_state["spoken_text"]:
        st.write(f"You said: {st.session_state['spoken_text']}")

    # Button to start recording
    if st.button("Start Recording"):
        recognized = recognize_speech()
        if recognized:
            st.session_state["spoken_text"] = recognized
            st.write(f"You said: {recognized}")

    # Button to predict
    if st.button("Predict"):
        if st.session_state["spoken_text"].strip():
            result = predict(st.session_state["spoken_text"])
            st.write(f"Predicted Sentiment: {result}")
        else:
            st.write("Please record some speech first.")