News Summarization and TTS Application

Introduction The News Summarization and Text-to-Speech (TTS) application is designed to extract, summarize, analyze sentiment, and convert news articles into speech. This project integrates web scraping, NLP-based text summarization, sentiment analysis, and text-to-speech conversion into a single web-based tool, allowing users to efficiently consume news content in multiple formats.

Features • News Extraction: Fetches articles using web scraping techniques. • Summarization: Generates concise summaries of lengthy news articles. • Sentiment Analysis: Determines the sentiment (positive, negative, or neutral) of summarized content. • Text-to-Speech Conversion: Converts text into speech, supporting Hindi and English. • User-Friendly Interface: A web-based UI built using Streamlit. • API Integration: Backend API for summarization, sentiment analysis, and text-to-speech functionalities.

Installation and Setup

Prerequisites Ensure you have the following installed on your system: • Python 3.8 or later • Git • Virtual environment (optional but recommended)

Steps to Set Up the Project 1. Clone the Repository git clone https://huggingface.co/spaces/riyaarahim/news-summarization-tts cd news-summarization-app 2. Create and Activate a Virtual Environment python -m venv venv source venv/bin/activate # macOS/Linux venv\Scripts\activate # Windows 3. Install Dependencies pip install -r requirements.txt 4. Run the Backend API python api.py 5. Run the Frontend Application streamlit run app.py 6. Deploying on Hugging Face Spaces ○ Ensure all dependencies are listed in requirements.txt. ○ Push the repository to Hugging Face. ○ Set app.py as the entry point. Dependencies The following key libraries and frameworks are used in this project: • Flask: For backend API development. • Streamlit: For building an interactive web application. • Newspaper3k: For extracting news content from websites. • Hugging Face Transformers: For NLP-based text summarization and sentiment analysis. • VADER (NLTK): A rule-based sentiment analysis tool. • gTTS: A Python library for text-to-speech conversion. Install all dependencies using: pip install -r requirements.txt

API Endpoints 1. Summarization API ○ Endpoint: /summarize ○ Method: POST ○ Input: { "text": "" } ○ Output: { "summary": "" } 2. Sentiment Analysis API ○ Endpoint: /sentiment ○ Method: POST ○ Input: { "text": "" } ○ Output: { "sentiment": "Positive/Negative/Neutral", "score": 0.85 } 3. Text-to-Speech API ○ Endpoint: /tts ○ Method: POST ○ Input: { "text": "", "language": "hi" } ○ Output: { "audio_url": "" }

Testing the API with Postman 1. Open Postman and create a new request. 2. Select POST and enter the endpoint URL. 3. In the Body section, select raw -> JSON. 4. Provide the required input JSON. 5. Click Send and verify the response.

Assumptions & Limitations Assumptions • Input news articles follow a structured format suitable for NLP processing. • A stable internet connection is required to fetch external models. • Summarization is optimized for longer texts. Limitations • The accuracy of summarization and sentiment analysis depends on the models used. • The TTS feature currently supports only Hindi and English. • Response times may vary based on text length and model complexity. • Some websites may block web scraping, limiting article extraction. Future Enhancements • Support for more languages in TTS. • Improved accuracy in summarization and sentiment analysis. • Real-time news updates and categorization. • Enhanced UI with additional interactive features.