app.py

import streamlit as st

from PIL import Image

from transformers import AutoTokenizer, AutoModelForSequenceClassification


# Define the model names or identifiers

model1_name = "Winnie-Kay/Sentiment-Analysis-Roberta-bases"
model2_name = "Winnie-Kay/Finetuned_BertModel_SentimentAnalysis"


# Initialize the tokenizer and models for sentiment analysis

tokenizer1 = AutoTokenizer.from_pretrained(model1_name)

model1 = AutoModelForSequenceClassification.from_pretrained(model1_name)

tokenizer2 = AutoTokenizer.from_pretrained(model2_name)

model2 = AutoModelForSequenceClassification.from_pretrained(model2_name)




# Define a function to preprocess the text data

def preprocess(text):

    new_text = []

    # Replace user mentions with '@user'

    for t in text.split(" "):

        t = '@user' if t.startswith('@') and len(t) > 1 else t

        # Replace links with 'http'

        t = 'http' if t.startswith('http') else t

        new_text.append(t)

    # Join the preprocessed text

    return " ".join(new_text)


# Define a function to perform sentiment analysis on the input text using model 1

def sentiment_analysis_model1(text):

    # Preprocess the input text

    text = preprocess(text)


    # Tokenize the input text using the pre-trained tokenizer

    encoded_input = tokenizer1(text, return_tensors='pt')


    # Feed the tokenized input to the pre-trained model and obtain output

    output = model1(**encoded_input)


    # Obtain the prediction scores for the output

    scores_ = output[0][0].detach().numpy()


    # Apply softmax activation function to obtain probability distribution over the labels

    scores_ = torch.nn.functional.softmax(torch.from_numpy(scores_), dim=0).numpy()


    # Format the output dictionary with the predicted scores

    labels = ['Negative', 'Positive']

    scores = {l:float(s) for (l,s) in zip(labels, scores_) }


    # Return the scores

    return scores


# Define a function to perform sentiment analysis on the input text using model 2

def sentiment_analysis_model2(text):

    # Preprocess the input text

    text = preprocess(text)


    # Tokenize the input text using the pre-trained tokenizer

    encoded_input = tokenizer2(text, return_tensors='pt')


    # Feed the tokenized input to the pre-trained model and obtain output

    output = model2(**encoded_input)


    # Obtain the prediction scores for the output

    scores_ = output[0][0].detach().numpy()


    # Apply softmax activation function to obtain probability distribution over the labels

    scores_ = torch.nn.functional.softmax(torch.from_numpy(scores_), dim=0).numpy()


    # Format the output dictionary with the predicted scores

    labels = ['Negative', 'Neutral', 'Positive']

    scores = {l:float(s) for (l,s) in zip(labels, scores_) }


    # Return the scores

    return scores


# Define the Streamlit app

def app():

    # Define the app title

    st.title("Sentiment Analysis")

    # Define the input field

    text_input = st.text_input("Enter text:")


    # Define the model selection dropdown

    model_selection = st.selectbox("Select a model:", ["Model 1", "Model 2"])


    # Perform sentiment analysis when the submit button is clicked

    if st.button("Submit"):

        if text_input:

            if model_selection == "Model 1":

                # Perform sentiment analysis using model 1

                scores = sentiment_analysis_model1(text_input)

                st.write(f"Model 1 predicted scores: {scores}")

            else:

                # Perform sentiment analysis using model 2

                scores = sentiment_analysis_model2(text_input)

                st.write(f"Model 2 predicted scores: {scores}")