app.py

import numpy as np
import pandas as pd
import re
import os
import cloudpickle
from transformers import (DebertaTokenizerFast, 
                          TFAutoModelForTokenClassification,
                          BartTokenizerFast, 
                          TFAutoModelForSeq2SeqLM)
import tensorflow as tf
import spacy
import streamlit as st
from scraper import scrape_text


os.environ['TF_USE_LEGACY_KERAS'] = "1"

class NERLabelEncoder:
    '''
    Label Encoder to encode and decode the entity labels
    '''
    def __init__(self):
        self.label_mapping = {'O': 0, 
                             'B-geo': 1, 
                             'I-geo': 2, 
                             'B-gpe': 3, 
                             'I-gpe': 4, 
                             'B-per': 5,
                             'I-per': 6,
                             'B-org': 7,
                             'I-org': 8,
                             'B-tim': 9,
                             'I-tim': 10,
                             'B-art': 11, 
                             'I-art': 12,
                             'B-nat': 13,
                             'I-nat': 14,
                             'B-eve': 15,
                             'I-eve': 16,
                             '[CLS]': -100,
                             '[SEP]': -100}
        
        self.inverse_label_mapping = {}
    
    def fit(self):
        self.inverse_label_mapping = {value: key for key, value in self.label_mapping.items()}
        return self
        
    def transform(self, x: pd.Series):
        x = x.map(self.label_mapping)
        return x
    
    def inverse_transform(self, x: pd.Series):
        x = x.map(self.inverse_label_mapping)
        return x


############ NER MODEL & VARS INITIALIZATION START ####################
NER_CHECKPOINT = "microsoft/deberta-base"
NER_N_TOKENS = 50
NER_N_LABELS = 18
NER_COLOR_MAP = {'GEO': '#DFFF00', 'GPE': '#FFBF00', 'PER': '#9FE2BF', 
                 'ORG': '#40E0D0', 'TIM': '#CCCCFF', 'ART': '#FFC0CB', 'NAT': '#FFE4B5', 'EVE': '#DCDCDC'}

@st.cache_resource
def load_ner_models():
    ner_model = TFAutoModelForTokenClassification.from_pretrained(NER_CHECKPOINT, num_labels=NER_N_LABELS, attention_probs_dropout_prob=0.4, hidden_dropout_prob=0.4)
    ner_model.load_weights(os.path.join("models", "general_ner_deberta_weights.h5"), by_name=True)
    ner_label_encoder = NERLabelEncoder()
    ner_label_encoder.fit()
    ner_tokenizer = DebertaTokenizerFast.from_pretrained(NER_CHECKPOINT, add_prefix_space=True)
    nlp = spacy.load(os.path.join('.', 'en_core_web_sm-3.6.0'))
    print('Loaded NER models')
    return ner_model, ner_label_encoder, ner_tokenizer, nlp

ner_model, ner_label_encoder, ner_tokenizer, nlp = load_ner_models()


############ NER MODEL & VARS INITIALIZATION END ####################

############ NER LOGIC START ####################
def softmax(x):
    return tf.exp(x) / tf.math.reduce_sum(tf.exp(x))

def ner_process_output(res):
    '''
    Function to concatenate sub-word tokens, labels and 
    compute mean prediction probability of tokens
    '''
    d = {}
    result = []
    pred_prob = []
    res.append(['-', 'B-b', 0])
    for n, i in enumerate(res):
        try:
            split = i[1].split('-')
            token = i[0]
            token_prob = i[2]
            prefix, suffix = split
            if prefix == 'B':
                if len(d) != 0:
                    result.append([(re.sub(r"[^\x00-\x7F]+", '', token.replace("Ġ", " ").strip()), label, np.mean(pred_prob))
                                   for label, token in d.items()][0])
                d = {}
                pred_prob = []
                pred_prob.append(token_prob)
                d[suffix] = token

            else:
                d[suffix] = d[suffix] + token
                pred_prob.append(token_prob)
        except:
            continue
            
    return result


def ner_inference(txt):
    '''
    Function that returns model prediction and prediction probabitliy
    '''
    test_data = [txt]
    # tokenizer = DebertaTokenizerFast.from_pretrained(NER_CHECKPOINT, add_prefix_space=True)
    tokens = ner_tokenizer.tokenize(txt)
    tokenized_data = ner_tokenizer(test_data, is_split_into_words=True, max_length=NER_N_TOKENS, 
                               truncation=True, padding="max_length")

    token_idx_to_consider = tokenized_data.word_ids()
    token_idx_to_consider = [i for i in range(len(token_idx_to_consider)) if token_idx_to_consider[i] is not None] 

    input_ = [tokenized_data['input_ids'], tokenized_data['attention_mask']]
    pred_logits = ner_model.predict(input_, verbose=0).logits[0]

    pred_prob = tf.map_fn(softmax, pred_logits)

    pred_idx = tf.argmax(pred_prob, axis=-1).numpy()
    pred_idx = pred_idx[token_idx_to_consider]

    pred_prob = tf.math.reduce_max(pred_prob, axis=-1).numpy()
    pred_prob = np.round(pred_prob[token_idx_to_consider], 3)
    pred_labels = ner_label_encoder.inverse_transform(pd.Series(pred_idx))

    result = [[token, label, prob] for token, label, 
              prob in zip(tokens, pred_labels, pred_prob) if label.find('-') >= 0]
    
    output = ner_process_output(result)
    return output


def ner_inference_long_text(txt):
    entities = []
    doc = nlp(txt)
    n_sents = len([_ for _ in doc.sents])
    n = 0
    progress_bar = st.progress(0, text=f'Processed 0 / {n_sents} sentences')
    for sent in doc.sents:
        entities.extend(ner_inference(sent.text))
        n += 1
        progress_bar.progress(n / n_sents, text=f'Processed {n} / {n_sents} sentences')
    # progress_bar.empty()
    return entities


def get_ner_text(article_txt, ner_result):
    res_txt = ''
    start = 0
    prev_start = 0
    for i in ner_result:
        try:
            span = next(re.finditer(fr'{i[0]}', article_txt)).span()
            start = span[0]
            end = span[1]
            res_txt += article_txt[prev_start:start]
            repl_str = f'''<span style="background-color:{NER_COLOR_MAP[i[1]]}; border-radius: 3px">{article_txt[start:end].strip()}
            <span style="font-size:10px; font-weight:bold; display:inline-block; vertical-align: middle;">
            {i[1]} ({str(np.round(i[2], 3))})</span></span>'''
            res_txt += article_txt[start:end].replace(article_txt[start:end], repl_str)
            prev_start = 0
            article_txt = article_txt[end:]
        except:
            continue
    res_txt += article_txt
    return res_txt

############ NER LOGIC END ####################


############ SUMMARIZATION MODEL & VARS INITIALIZATION START ####################
SUMM_CHECKPOINT = "facebook/bart-base"
SUMM_INPUT_N_TOKENS = 400
SUMM_TARGET_N_TOKENS = 300

@st.cache_resource
def load_summarizer_models():
    summ_tokenizer = BartTokenizerFast.from_pretrained(SUMM_CHECKPOINT)
    summ_model = TFAutoModelForSeq2SeqLM.from_pretrained(SUMM_CHECKPOINT)
    summ_model.load_weights(os.path.join("models", "bart_en_summarizer.h5"), by_name=True)
    print('Loaded summarizer models')
    return summ_tokenizer, summ_model

summ_tokenizer, summ_model = load_summarizer_models()

def summ_preprocess(txt):
    txt = re.sub(r'^By \. [\w\s]+ \. ', ' ', txt) # By . Ellie Zolfagharifard . 
    txt = re.sub(r'\d{1,2}\:\d\d [a-zA-Z]{3}', ' ', txt) # 10:30 EST
    txt = re.sub(r'\d{1,2} [a-zA-Z]+ \d{4}', ' ', txt) # 10 November 1990
    txt = txt.replace('PUBLISHED:', ' ')
    txt = txt.replace('UPDATED', ' ')
    txt = re.sub(r' [\,\.\:\'\;\|] ', ' ', txt) # remove puncts with spaces before and after
    txt = txt.replace(' : ', ' ')
    txt = txt.replace('(CNN)', ' ')
    txt = txt.replace('--', ' ')
    txt = re.sub(r'^\s*[\,\.\:\'\;\|]', ' ', txt) # remove puncts at beginning of sent
    txt = re.sub(r' [\,\.\:\'\;\|] ', ' ', txt) # remove puncts with spaces before and after
    txt = re.sub(r'\n+',' ', txt)
    txt = " ".join(txt.split())
    return txt

def summ_inference_tokenize(input_: list, n_tokens: int):
    tokenized_data = summ_tokenizer(text=input_, max_length=SUMM_TARGET_N_TOKENS, truncation=True, padding="max_length", return_tensors="tf")
    return summ_tokenizer, tokenized_data    

def clean_summary(summary: str):
    summary = summary.strip()
    if summary[-1] != '.':
        sents = summary.split(". ")
        summary = ". ".join(sents[:-1])
        summary += "."
    summary = re.sub(r'^-', "", summary)
    summary = summary.strip()
    if len(summary) <= 5:
        summary = ""
    return summary

def summ_inference(txt: str):
    txt = summ_preprocess(txt)
    inference_tokenizer, tokenized_data = summ_inference_tokenize(input_=[txt], n_tokens=SUMM_INPUT_N_TOKENS)
    pred = summ_model.generate(**tokenized_data, max_new_tokens=SUMM_TARGET_N_TOKENS)
    result = "" if txt=="" else clean_summary(inference_tokenizer.decode(pred[0], skip_special_tokens=True))
    return result
############ SUMMARIZATION MODEL & VARS INITIALIZATION END ####################

############## ENTRY POINT START #######################
def main():
    st.markdown('''<h3>News Summarizer and NER</h3> 
    <p><a href="https://huggingface.co/spaces/ksvmuralidhar/news_summarizer_ner/blob/main/README.md#new-summarization-and-ner" target="_blank">README</a>
    <br>
    The app works best in summarizing <a href="https://edition.cnn.com/" target="_blank">CNN</a> and 
    <a href="https://www.dailymail.co.uk/home/index.html" target="_blank">Daily Mail</a> news articles, 
    as the BART model is fine-tuned on them.
    </p>
    
    ''', unsafe_allow_html=True)
    input_type = st.radio('Select an option:', ['Paste news URL', 'Paste news text'], 
                      horizontal=True)

    scrape_error = None
    summary_error = None
    ner_error = None
    summ_result = None
    ner_result = None
    ner_df = None
    article_txt = None
    
    
    if input_type == 'Paste news URL':
        article_url = st.text_input("Paste the URL of a news article", "")
        
        if (st.button("Submit")) or (article_url):
            with st.status("Processing...", expanded=True) as status:
                status.empty()
                # Scraping data Start
                try:
                    st.info("Scraping data from the URL.", icon="ℹ️")
                    article_txt = scrape_text(article_url)
                    st.success("Successfully scraped the data.", icon="✅")
                except Exception as e:
                    article_txt = None
                    scrape_error = str(e) 

                # Scraping data End
    
                if article_txt is not None:
                    article_txt = re.sub(r'\n+',' ', article_txt)

                    # Generating summary start
                    
                    try:
                        st.info("Generating the summary.", icon="ℹ️")
                        summ_result = summ_inference(article_txt)
                    except Exception as e:
                        summ_result = None
                        summary_error = str(e)
                    if summ_result is not None:
                        st.success("Successfully generated the summary.", icon="✅")
                    else:
                        st.error("Encountered an error while generating the summary.", icon="🚨")

                    # Generating summary end

                    
                    # NER start
                    try:
                        st.info("Recognizing the entites.", icon="ℹ️")
                        ner_result = [[ent, label.upper(), np.round(prob, 3)] 
                                      for ent, label, prob in ner_inference_long_text(article_txt)]

                        ner_df = pd.DataFrame(ner_result, columns=['entity', 'label', 'confidence'])

                        ner_result = get_ner_text(article_txt, ner_result).replace('$', '\$')
            
                    except Exception as e:
                        ner_result = None
                        ner_error = str(e)
                    if ner_result is not None:
                        st.success("Successfully recognized the entites.", icon="✅")
                    else:
                        st.error("Encountered an error while recognizing the entites.", icon="🚨")

                    # NER end                                 
                else:
                    st.error("Encountered an error while scraping the data.", icon="🚨")

                if (scrape_error is None) and (summary_error is None) and (ner_error is None):
                    status.update(label="Done", state="complete", expanded=False)
                else:
                    status.update(label="Error", state="error", expanded=False)

            if scrape_error is not None:
                st.error(f"Scrape Error:  \n{scrape_error}", icon="🚨")
            else:
                if summary_error is not None:
                    st.error(f"Summary Error:  \n{summary_error}", icon="🚨")
                else:
                    st.markdown(f"<h4>SUMMARY:</h4>{summ_result}", unsafe_allow_html=True)
                    
                if ner_error is not None:
                    st.error(f"NER Error  \n{ner_error}", icon="🚨")
                else:
                    st.markdown(f"<h4>ENTITIES:</h4>{ner_result}", unsafe_allow_html=True)
                    # st.dataframe(ner_df, use_container_width=True)
    
                st.markdown(f"<h4>SCRAPED TEXT:</h4>{article_txt}", unsafe_allow_html=True)       

    else:
        article_txt = st.text_area("Paste the text of a news article", "", height=150)

        if (st.button("Submit")) or (article_txt):
            with st.status("Processing...", expanded=True) as status:
                article_txt = re.sub(r'\n+',' ', article_txt)

                # Generating summary start
                
                try:
                    st.info("Generating the summary.", icon="ℹ️")
                    summ_result = summ_inference(article_txt)
                except Exception as e:
                    summ_result = None
                    summary_error = str(e)
                if summ_result is not None:
                    st.success("Successfully generated the summary.", icon="✅")
                else:
                    st.error("Encountered an error while generating the summary.", icon="🚨")

                # Generating summary end

                
                # NER start
                try:
                    st.info("Recognizing the entites.", icon="ℹ️")
                    ner_result = [[ent, label.upper(), np.round(prob, 3)] 
                                  for ent, label, prob in ner_inference_long_text(article_txt)]

                    ner_df = pd.DataFrame(ner_result, columns=['entity', 'label', 'confidence'])

                    ner_result = get_ner_text(article_txt, ner_result).replace('$', '\$')
        
                except Exception as e:
                    ner_result = None
                    ner_error = str(e)
                if ner_result is not None:
                    st.success("Successfully recognized the entites.", icon="✅")
                else:
                    st.error("Encountered an error while recognizing the entites.", icon="🚨")

                    # NER end                                 

                if (summary_error is None) and (ner_error is None):
                    status.update(label="Done", state="complete", expanded=False)
                else:
                    status.update(label="Error", state="error", expanded=False)

            if summary_error is not None:
                st.error(f"Summary Error:  \n{summary_error}", icon="🚨")
            else:
                st.markdown(f"<h4>SUMMARY:</h4>{summ_result}", unsafe_allow_html=True)
                
            if ner_error is not None:
                st.error(f"NER Error  \n{ner_error}", icon="🚨")
            else:
                st.markdown(f"<h4>ENTITIES:</h4>{ner_result}", unsafe_allow_html=True)
                # st.dataframe(ner_df, use_container_width=True)
        
        

############## ENTRY POINT END #######################

if __name__ == "__main__":
    main()