app.py

import streamlit as st
from transformers import pipeline
import math
import nltk
from nltk.corpus import stopwords
nltk.download('punkt')
nltk.download('stopwords')
sentiment_model = pipeline("text-classification", model="AhmedTaha012/managersFeedback-V1.0.7")
increase_decrease_model = pipeline("text-classification", model="AhmedTaha012/nextQuarter-status-V1.1.9")
ner_model = pipeline("token-classification", model="AhmedTaha012/finance-ner-v0.0.9-finetuned-ner")
def getSpeakers(data):
    if "Speakers" in data:
        return "\n".join([x for x in data.split("Speakers")[-1].split("\n") if "--" in x])
    elif "Call participants" in data:
        return "\n".join([x for x in data.split("Call participants")[-1].split("\n") if "--" in x])
    elif "Call Participants" in data:
        return "\n".join([x for x in data.split("Call Participants")[-1].split("\n") if "--" in x])
def removeSpeakers(data):
    if "Speakers" in data:
        return data.split("Speakers")[0]
    elif "Call participants" in data:
        return data.split("Call participants")[0]
    elif "Call Participants" in data:
        return data.split("Call Participants")[0]
def getQA(data):
    if "Questions and Answers" in data:    
        return data.split("Questions and Answers")[-1]
    elif  "Questions & Answers" in data:
        return data.split("Questions & Answers")[-1]
    elif "Q&A" in data:
        return data.split("Q&A")[-1]
    else:
        return ""
def removeQA(data):
    if "Questions and Answers" in data:    
        return data.split("Questions and Answers")[0]
    elif  "Questions & Answers" in data:
        return data.split("Questions & Answers")[0]
    elif "Q&A" in data:
        return data.split("Q&A")[0]
    else:
        return ""
def clean_and_preprocess(text):
    text=[x for x in text.split("\n") if len(x)>100]
    l=[]
    for t in text:
        # Convert to lowercase
        t = t.lower()    
        # Tokenize text into words
        words = nltk.word_tokenize(t)
        # Remove stopwords
        stop_words = set(stopwords.words('english'))
        filtered_words = [word for word in words if word not in stop_words]

        # Join the words back into a cleaned text
        cleaned_text = ' '.join(filtered_words)
        l.append(cleaned_text)
    return "\n".join(l)
def replace_abbreviations(text):
    
    replacements = {
        'Q1': 'first quarter',
        'Q2': 'second quarter',
        'Q3': 'third quarter',
        'Q4': 'fourth quarter',
        'q1': 'first quarter',
        'q2': 'second quarter',
        'q3': 'third quarter',
        'q4': 'fourth quarter',
        'FY': 'fiscal year',
        'YoY': 'year over year',
        'MoM': 'month over month',
        'EBITDA': 'earnings before interest, taxes, depreciation, and amortization',
        'ROI': 'return on investment',
        'EPS': 'earnings per share',
        'P/E': 'price-to-earnings',
        'DCF': 'discounted cash flow',
        'CAGR': 'compound annual growth rate',
        'GDP': 'gross domestic product',
        'CFO': 'chief financial officer',
        'GAAP': 'generally accepted accounting principles',
        'SEC': 'U.S. Securities and Exchange Commission',
        'IPO': 'initial public offering',
        'M&A': 'mergers and acquisitions',
        'EBIT': 'earnings before interest and taxes',
        'IRR': 'internal rate of return',
        'ROA': 'return on assets',
        'ROE': 'return on equity',
        'NAV': 'net asset value',
        'PE ratio': 'price-to-earnings ratio',
        'EPS growth': 'earnings per share growth',
        'Fiscal Year': 'financial year',
        'CAPEX': 'capital expenditure',
        'APR': 'annual percentage rate',
        'P&L': 'profit and loss',
        'NPM': 'net profit margin',
        'EBT': 'earnings before taxes',
        'EBITDAR': 'earnings before interest, taxes, depreciation, amortization, and rent',
        'PAT': 'profit after tax',
        'COGS': 'cost of goods sold',
        'EBTIDA': 'earnings before taxes, interest, depreciation, and amortization',
        'E&Y': 'Ernst & Young',
        'B2B': 'business to business',
        'B2C': 'business to consumer',
        'LIFO': 'last in, first out',
        'FIFO': 'first in, first out',
        'FCF': 'free cash flow',
        'LTM': 'last twelve months',
        'OPEX': 'operating expenses',
        'TSR': 'total shareholder return',
        'PP&E': 'property, plant, and equipment',
        'PBT': 'profit before tax',
        'EBITDAR margin': 'earnings before interest, taxes, depreciation, amortization, and rent margin',
        'ROIC': 'return on invested capital',
        'EPS': 'earnings per share',
    'P/E': 'price-to-earnings',
    'EBITDA': 'earnings before interest, taxes, depreciation, and amortization',
    'YOY': 'year-over-year',
    'MOM': 'month-over-month',
    'CAGR': 'compound annual growth rate',
    'GDP': 'gross domestic product',
    'ROI': 'return on investment',
    'ROE': 'return on equity',
    'EBIT': 'earnings before interest and taxes',
    'DCF': 'discounted cash flow',
    'GAAP': 'Generally Accepted Accounting Principles',
    'LTM': 'last twelve months',
    'EBIT margin': 'earnings before interest and taxes margin',
    'EBT': 'earnings before taxes',
    'EBTA': 'earnings before taxes and amortization',
    'FTE': 'full-time equivalent',
    'EBIDTA': 'earnings before interest, depreciation, taxes, and amortization',
    'EBTIDA': 'earnings before taxes, interest, depreciation, and amortization',
    'EBITDAR': 'earnings before interest, taxes, depreciation, amortization, and rent',
    'COGS': 'cost of goods sold',
    'APR': 'annual percentage rate',
    'PESTEL': 'Political, Economic, Social, Technological, Environmental, and Legal',
    'KPI': 'key performance indicator',
    'SWOT': 'Strengths, Weaknesses, Opportunities, Threats',
    'CAPEX': 'capital expenditures',
    'EBITDARM': 'earnings before interest, taxes, depreciation, amortization, rent, and management fees',
    'EBITDAX': 'earnings before interest, taxes, depreciation, amortization, and exploration expenses',
    'EBITDAS': 'earnings before interest, taxes, depreciation, amortization, and restructuring costs',
    'EBITDAX-C': 'earnings before interest, taxes, depreciation, amortization, exploration expenses, and commodity derivatives',
    'EBITDAX-R': 'earnings before interest, taxes, depreciation, amortization, exploration expenses, and asset retirement obligations',
    'EBITDAX-E': 'earnings before interest, taxes, depreciation, amortization, exploration expenses, and environmental liabilities'
        
        # Add more abbreviations and replacements as needed
    }
    for abbreviation, full_form in replacements.items():
        text = text.replace(abbreviation, full_form)
    
    return text

def clean_and_preprocess(text):
    text=[x for x in text.split("\n") if len(x)>100]
    l=[]
    for t in text:
        # Convert to lowercase
        t = t.lower()    
        # Tokenize text into words
        words = nltk.word_tokenize(t)
        # Remove stopwords
        stop_words = set(stopwords.words('english'))
        filtered_words = [word for word in words if word not in stop_words]

        # Join the words back into a cleaned text
        cleaned_text = ' '.join(filtered_words)
        l.append(cleaned_text)
    return "\n".join(l)







st.title("Transcript Analysis")
transcript = st.text_area("Enter the transcript:", height=200)
transcript=replace_abbreviations(transcript)
transcript=replace_abbreviations(transcript)
transcript=removeSpeakers(transcript)
transcript=removeQA(transcript)
transcript=clean_and_preprocess(transcript)
tokens=transcript.split()
splitSize=256
chunks=[tokens[r*splitSize:(r+1)*splitSize] for r in range(math.ceil(len(tokens)/splitSize))]

if st.button("Analyze"):
    st.subheader("Sentiment Analysis")
    sentiment = [sentiment_model(x)[0]['label'] for x in chunks]
    sentiment=max(sentiment,key=sentiment.count)
    sentiment_color = "green" if sentiment == "POSITIVE" else "red"
    st.markdown(f'<span style="color:{sentiment_color}">{sentiment}</span>', unsafe_allow_html=True)

    st.subheader("Increase/Decrease Prediction")
    increase_decrease = [increase_decrease_model(x)[0]['label'] for x in chunks]
    increase_decrease=max(increase_decrease,key=increase_decrease.count)
    increase_decrease_color = "green" if increase_decrease == "INCREASE" else "red"
    st.markdown(f'<span style="color:{increase_decrease_color}">{increase_decrease}</span>', unsafe_allow_html=True)

    st.subheader("NER Metrics")
    ner_result = [ner_model(x) for x in chunks]
    st.write(str(ner_result))