app.py

import requests
import streamlit as st
import torch
import joblib
import tensorflow as tf
from transformers import AutoTokenizer, LEDForConditionalGeneration
from tensorflow.keras.models import load_model
from transformers import TFBertForSequenceClassification, BertTokenizer
st.set_page_config(page_title="Summarization&tweet_analysis", page_icon="📈",layout="wide")
hide_streamlit_style = """
            <style>
            #MainMenu {visibility: hidden;}
            footer {	
	visibility: hidden;
	
	}
footer:after {
	content:'©2023 Sravathi AI Technology. All rights reserved.'; 
	visibility: visible;
	display: block;
	position: relative;
	#background-color: red;
	padding: 5px;
	top: 2px;
}
            </style>
            """
st.markdown(hide_streamlit_style, unsafe_allow_html=True)
import pandas as pd
import time
import sys
import pickle
#from stqdm import stqdm
import base64
#from tensorflow.keras.preprocessing.text import Tokenizer
#from tensorflow.keras.preprocessing.sequence import pad_sequences
import numpy as np
import json
import os
import re

#from tensorflow.keras.models import load_model
#st.write("API examples - Dermatophagoides, Miconazole, neomycin,Iothalamate")

#background_image = sys.path[1]+"/streamlit_datafile_links/audience-1853662_960_720.jpg"  # Path to your background image
def add_bg_from_local(image_file):
    with open(image_file, "rb") as image_file:
        encoded_string = base64.b64encode(image_file.read())
    st.markdown(
    f"""
    <style>
    .stApp {{
        background-image: url(data:image/{"jpg"};base64,{encoded_string.decode()});
        background-size: cover
    }}
    </style>
    """,
    unsafe_allow_html=True
    )
#add_bg_from_local(background_image)  

#@st.cache
st.header('Summarization & tweet_analysis')
def convert_df(df):
                # IMPORTANT: Cache the conversion to prevent computation on every rerun
                return df.to_csv(index=False).encode('utf-8')  
col1, col2 = st.columns([4,1])
result_csv_batch_sql = result_csv_batch_fail=result_csv_batch=result_csv4=result_csv3=result_csv1=result_csv2=0

with col1:
    models = st.selectbox(
    'Select the option',
    ('summarization_model1','tweet_analysis' ))
    #try: 
    if models == 'summarization_model1':   
        st.markdown("")
    else:
         st.markdown("")
    with st.form("form1"):
        hide_label = """
        <style>
            .css-9ycgxx {
                display: none;
            }
        </style>
        """
        text_data = st.text_input('Enter the text')
        print(text_data)
        st.markdown(hide_label, unsafe_allow_html=True)            
        submitted = st.form_submit_button("Submit")
        if submitted:
            if models == 'summarization_model1': 
                #torch.cuda.set_device(2)
                tokenizer = AutoTokenizer.from_pretrained('allenai/PRIMERA-multinews')
                model = LEDForConditionalGeneration.from_pretrained('allenai/PRIMERA-multinews')
                #device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') # get the device
                device = "cpu"
                model.to(device) # move the model to the device  
                documents = text_data
    
                # Tokenize and encode the documents
                inputs = tokenizer(documents, return_tensors='pt', padding=True, truncation=True,max_length=1000000)
    
                # Move the inputs to the device
                inputs = inputs.to(device)
    
                # Generate summaries
                outputs = model.generate(**inputs,max_length=1000000)
    
                # Decode the summaries
                st.write(tokenizer.batch_decode(outputs, skip_special_tokens=True))                          
                st.success('Prediction done successfully!', icon="✅")  
            else:
                    # Define the custom objects (custom layers) needed for loading the model
                    custom_objects = {"TFBertForSequenceClassification": TFBertForSequenceClassification}
                    
                    # Load the best model checkpoint
                    best_model = load_model('best_model_checkpoint_val_acc_0.8697_epoch_03.h5', custom_objects=custom_objects)

                    # Assuming you already have the test set DataFrame (df_test) and tokenizer
                    tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
                    test_encodings = tokenizer(text_data, padding=True, truncation=True, return_tensors='tf')
                    test_dataset = tf.data.Dataset.from_tensor_slices((dict(test_encodings)))

                    # Make predictions on the test set using the loaded model
                    predictions_probabilities = best_model.predict(test_dataset.batch(8))

                    # Convert probabilities to one-hot encoded predictions
                    predictions_onehot = np.eye(9)[np.argmax(predictions_probabilities, axis=1)]

                    # Display or save the DataFrame with predicted labels
                    index_arg = np.argmax(predictions_probabilities, axis=1)
                    # Later, you can load the LabelEncoder
                    label_encoder = joblib.load('label_encoder.joblib')
                    result_label = label_encoder.inverse_transform(index_arg)

                    # Display or save the DataFrame with predicted labels
                    st.write("Item name: ", result_label[0])
                    
                    from transformers import AutoTokenizer, AutoConfig, AutoModelForSequenceClassification
                    from scipy.special import softmax

                    MODEL = f"cardiffnlp/twitter-roberta-base-sentiment-latest"
                    tokenizer = AutoTokenizer.from_pretrained(MODEL)
                    config = AutoConfig.from_pretrained(MODEL)
                    # PT
                    model = AutoModelForSequenceClassification.from_pretrained(MODEL)
                    #model.save_pretrained(MODEL)
                    #text = "Covid cases are increasing fast!"
                    pred_label = []
                    pred_scor = []
                    def preprocess(text):
                        new_text = []
                        for t in text.split(" "):
                            t = '@user' if t.startswith('@') and len(t) > 1 else t
                            t = 'http' if t.startswith('http') else t
                            new_text.append(t)
                        return " ".join(new_text)
                    def predict_pret(text):
                        #print(text)
                        text = preprocess(text)
                        encoded_input = tokenizer(text, return_tensors='pt')
                        output = model(**encoded_input)
                        scores = output[0][0].detach().numpy()
                        scores = softmax(scores)

                        ranking = np.argsort(scores)
                        ranking = ranking[::-1]
                        l = config.id2label[ranking[0]]
                        s = scores[ranking[0]]
                        return l,s

                    l,s = predict_pret(text_data)

                    st.write("Sentiment is: ", l)

                    st.success('Prediction done successfully!', icon="✅")                 
        _='''
        except Exception as e:
            if 'NoneType' or 'not defined' in str(e):
                st.warning('Enter the required inputs', icon="⚠️")
            else:          
                st.warning(str(e), icon="⚠️")
        ''' 

for i in range(30):
    st.markdown('##')