Introduction
SARAHAI-SENTv1.2 is a refined single-file Python application for sentiment analysis, inspired by the technologies and algorithms outlined in U.S. Patent No. 11,308,384. It provides:
Autonomous Data Gathering via basic web scraping.
Easy Text Input (both via URL and manual entry).
Sentiment Analysis using a PyTorch-based model, removing TensorFlow/Keras dependencies.
User-Friendly Web Interface (Flask-based).
Interactive Visualizations (Plotly charts).
OpenDocument Report Generation (ODT, ODS).
This guide details each feature and explains how you can install, run, and use SARAHAI-SENTv1.2 to analyze sentiment on Windows 11 (or any compatible OS) without worrying about Keras 3/TensorFlow issues.
2. Key Features
2.1 Autonomous Data Gathering
Web Scraping: Input a URL, and SARAHAI-SENTv1.2 fetches the page’s HTML to extract visible text for analysis.
Basic Implementation: Uses the requests and BeautifulSoup libraries to gather text content.
Extendable: Can be enhanced to scrape multiple URLs, follow links, or parse structured formats (JSON, XML, etc.).
2.2 Sentiment Analysis (PyTorch-Only)
Pre-trained Model: Default is distilbert-base-uncased-finetuned-sst-2-english from Hugging Face.
No TensorFlow/Keras: Explicitly uses PyTorch to avoid incompatibility with Keras 3.
Transformer Pipeline: Utilizes the Hugging Face pipeline for straightforward sentiment classification.