app.py

import pybase64 as base64
import pandas as pd
import streamlit as st
import seaborn as sns
from data_cleaning import preprocess
from transformers import pipeline
from data_integration import scrape_all_pages

#@st.cache_data
#def get_img_as_base64(file):
#    with open(file, "rb") as f:
#        data = f.read()
#    return base64.b64encode(data).decode()


#img = get_img_as_base64("image.jpg")background-image: url("data:image/png;base64,{img}");


page_bg_img = """
<style>
.stApp > header {
    background-color: transparent;
}

.stApp {

    background-color:hsla(244,100%,50%,1);
    background-image:
    radial-gradient(at 40% 20%, hsla(266,100%,49%,1) 0px, transparent 50%),
    radial-gradient(at 80% 0%, hsla(189,100%,56%,1) 0px, transparent 50%);
    background-size: 150% 150%;
    animation: my_animation 10s ease infinite;
}

@keyframes my_animation {
    0% {
        background-position: 0% 0%;
    }
    25% {
        background-position: 100% 0%;
    }
    50% {
        background-position: 100% 100%;
    }
    75% {
        background-position: 0% 100%;
    }
    100% {
        background-position: 0% 0%;
    }
}

</style>
"""
st.markdown(page_bg_img, unsafe_allow_html=True)
#st.image("logo.png", width=200, height=200)

st.image("logo.png", width=80)
st.subheader(':violet[NLP HUB®]')
st.markdown("")
st.markdown("")
st.markdown("")
st.markdown("")

st.subheader('Amazon Sentiment Analysis using FineTuned :red[GPT-2] Pre-Trained Model')

sentiment_model = pipeline(model="ashok2216/gpt2-amazon-sentiment-classifier")
# Example usage:-
sample_url = 'https://www.amazon.in/Dell-Inspiron-i7-1255U-Processor-Platinum/product-reviews/B0C9F142V6/ref=cm_cr_dp_d_show_all_btm?ie=UTF8&reviewerType=all_reviews'
url = st.text_input("Amazon product link", sample_url)
st.button("Re-run")
st.write("Done")
st.subheader('', divider='rainbow')

try:
    all_reviews = scrape_all_pages(url)
    # Convert to DataFrame for further analysis
    reviews = pd.DataFrame(all_reviews)
    reviews['processed_text'] = reviews['content'].apply(preprocess)
    
    # st.dataframe(reviews, use_container_width=True)
    # st.markdown(sentiment_model(['It is Super!']))
    
    sentiments = []
    for text in reviews['processed_text']:
        if list(sentiment_model(text)[0].values())[0] == 'LABEL_1':
            output = 'Positive'
        else:
            output = 'Negative'
        sentiments.append(output)
    
    reviews['sentiments'] = sentiments
    st.markdown(':white[Output]')
    st.dataframe(reviews, use_container_width=True)
    # sns.countplot(reviews['sentiments'])
except KeyError:
    st.markdown('Please :red[Re-run] the app')