app.py

import os
import streamlit as st
import praw
import googleapiclient.discovery
import joblib
import numpy as np
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LinearRegression
from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer
from transformers import pipeline

# Load environment variables
REDDIT_CLIENT_ID = os.getenv("REDDIT_CLIENT_ID")
REDDIT_CLIENT_SECRET = os.getenv("REDDIT_CLIENT_SECRET")
REDDIT_USER_AGENT = os.getenv("REDDIT_USER_AGENT")
YOUTUBE_API_KEY = os.getenv("YOUTUBE_API_KEY")

# Authenticate Reddit
def authenticate_reddit():
    return praw.Reddit(
        client_id=REDDIT_CLIENT_ID,
        client_secret=REDDIT_CLIENT_SECRET,
        user_agent=REDDIT_USER_AGENT
    )

# Authenticate YouTube
def authenticate_youtube():
    return googleapiclient.discovery.build("youtube", "v3", developerKey=YOUTUBE_API_KEY)

# VADER Sentiment Analysis
vader = SentimentIntensityAnalyzer()

def get_vader_sentiment(text):
    scores = vader.polarity_scores(text)
    return scores['compound']  # Ranges from -1 (negative) to +1 (positive)

# BERT Sentiment Analysis
bert_sentiment = pipeline("sentiment-analysis")

def get_bert_sentiment(text):
    result = bert_sentiment(text)[0]
    return result['label'], result['score']

# Regression Sentiment Analysis
vectorizer = TfidfVectorizer()
regressor = LinearRegression()

def train_regression_model():
    sample_data = [
        ("I love this!", 1.0),
        ("This is amazing", 0.9),
        ("It's okay", 0.5),
        ("Not great", 0.3),
        ("I hate this", 0.1)
    ]
    texts, scores = zip(*sample_data)
    X = vectorizer.fit_transform(texts)
    regressor.fit(X, scores)
    joblib.dump((vectorizer, regressor), "sentiment_model.pkl")

train_regression_model()

# Predict with Regression Model
def get_regression_sentiment(text):
    vectorizer, regressor = joblib.load("sentiment_model.pkl")
    X = vectorizer.transform([text])
    return regressor.predict(X)[0]

# Streamlit UI
st.title("Sentiment Analysis App")
user_input = st.text_area("Enter text for sentiment analysis")
if st.button("Analyze"):
    vader_score = get_vader_sentiment(user_input)
    bert_label, bert_score = get_bert_sentiment(user_input)
    regression_score = get_regression_sentiment(user_input)
    
    st.write(f"**VADER Sentiment Score:** {vader_score}")
    st.write(f"**BERT Sentiment:** {bert_label} ({bert_score:.2f})")
    st.write(f"**Regression Sentiment Score:** {regression_score:.2f}")