A powerful and user-friendly command-line tool for analyzing the sentiment of text using state-of-the-art transformer models.

• 🧠 Powerful NLP: Leverages Hugging Face transformer models for accurate sentiment analysis
• 🚀 High Performance: Automatic GPU acceleration when available
• 📊 Data Visualization: Generate charts and statistics from your analyses
• 🌈 Rich Output: Color-coded results with visual confidence indicators
• 📝 Comprehensive Logging: Track all analyses in a structured CSV format
• 📦 Batch Processing: Analyze multiple texts from files for efficiency
• 🛠️ Flexible Usage: Interactive mode or command-line options

• Python 3.7+
• pip (Python package manager)

1. Clone this repository:

   git clone https://github.com/yourusername/sentiment-analysis-tool.git
   cd sentiment-analysis-tool

2. Install the required dependencies:

   pip install -r requirements.txt

   Or install them directly:

   pip install transformers torch tqdm colorama matplotlib

Simply run the script without arguments for interactive mode:

python sentiment_analyzer.py

Enter text at the prompt to analyze sentiment. Special commands:

• stats: Display statistics about your session
• viz: Generate and display a visualization of sentiment distribution
• quit: Exit the program

python sentiment_analyzer.py --file input.txt --output results.csv --model cardiffnlp/twitter-roberta-base-sentiment

Available options:

• --model: Specify the Hugging Face model to use (default: "cardiffnlp/twitter-roberta-base-sentiment")
• --file: Process multiple texts from a file (one per line)
• --output: Save results to a specified CSV file
• --no-log: Disable automatic logging of results
• --cpu: Force CPU usage even if GPU is available

python sentiment_analyzer.py
> I absolutely love this new feature, it works perfectly!

python sentiment_analyzer.py --file customer_reviews.txt --output sentiment_results.csv

python sentiment_analyzer.py --model distilbert-base-uncased-finetuned-sst-2-english

You can import and use the SentimentAnalyzer class in your own Python code:

from sentiment_analyzer import SentimentAnalyzer

analyzer = SentimentAnalyzer()
result = analyzer.analyze("I'm really enjoying this new software!")
print(f"Sentiment: {result['sentiment']}, Score: {result['score']}")

texts = ["Great product!", "Not satisfied with the quality", "It's okay, but not amazing"]
results = analyzer.analyze_batch(texts)

By default, all analyses are logged to sentiment_log.csv with timestamps, which allows for:

• Tracking sentiment trends over time
• Building datasets for further analysis
• Generating comprehensive reports

The tool can generate pie charts showing the distribution of sentiments in your analyzed texts:

• transformers: State-of-the-art NLP models
• torch: Deep learning framework
• tqdm: Progress bars for batch processing
• colorama: Terminal text coloring
• matplotlib: Data visualization

Contributions are welcome!

• Hugging Face for their excellent transformers library
• Cardiff NLP for the pre-trained sentiment analysis model

Distributed under the GNU Affero General Public License v3.0 License. See LICENSE for more information.