webrover 0.1.4

Generate high-quality datasets from web content for AI training

WebRover 🚀

WebRover is a powerful Python library for generating high-quality datasets from web content, designed specifically for training Large Language Models and AI applications.

---

🌟 Features

• Smart Web Scraping: Automatically find and scrape relevant content based on topics
• Multiple Input Formats: Support for JSON, YAML, TXT, and Markdown topic files
• Async Processing: Fast, concurrent scraping with built-in rate limiting
• Quality Control: Built-in content validation and cleaning
• LLM-Ready Output: Structured JSONL format perfect for model training
• Error Handling: Robust error tracking and recovery mechanisms

🚀 Quick Start

Installation

pip install webrover

Basic Usage

from webrover import WebRover

# Initialize WebRover
rover = WebRover()

# Scrape content from topics
rover.scrape_topics(
    topics=["artificial intelligence", "machine learning"],
    num_websites=100
)

# Save the dataset
rover.save_dataset("my_dataset.jsonl")

Using Topic Files

# From JSON file
rover.scrape_topics(
    topics="topics.json",
    num_websites=100
)

# From Markdown list
rover.scrape_topics(
    topics="topics.md",
    num_websites=100
)

📖 Documentation

Supported Topic File Formats

JSON

{
    "topics": [
        "AI basics",
        "machine learning",
        "deep learning"
    ]
}

YAML

topics:
  - AI basics
  - machine learning
  - deep learning

Markdown

- AI basics
- machine learning
- deep learning

Output Structure

{
    'url': 'https://example.com/article',
    'title': 'Article Title',
    'content': 'Article content...',
    'metadata': {
        'length': 1234,
        'has_title': true,
        'domain': 'example.com'
    }
}

🛠️ Advanced Usage

# Initialize with custom output directory
rover = WebRover(output_dir="my_datasets")

# Get scraping statistics
stats = rover.get_stats()
print(f"Success rate: {stats['success_rate']*100:.1f}%")

# Access dataset programmatically
dataset = rover.get_dataset()

📊 Output Files

• final_dataset/dataset.jsonl: Main dataset in JSONL format
• websites_master.json: List of all discovered URLs
• websites_completed.json: Successfully scraped URLs
• websites_errors.json: Failed attempts with error details

🔄 Error Handling

WebRover automatically handles common issues:

• Rate limiting
• Network timeouts
• Invalid URLs
• Blocked requests
• Malformed content

🚧 Limitations

• Respects robots.txt and site rate limits
• Some sites may block automated access
• Large datasets require more processing time
• Google search may throttle excessive requests

🗺️ Roadmap

See FUTURE.md for planned features and improvements.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

Built with ❤️ by Area-25. Special thanks to all contributors.

---

WebRover: Build better datasets, train better models. 🚀