reasoning-gym/python

"""Generate a markdown gallery of all available datasets with examples"""

from pathlib import Path
import textwrap

from reasoning_gym.factory import DATASETS, create_dataset


def generate_gallery() -> str:
    """Generate markdown content for the gallery"""
    
    # Start with header
    content = ["# Dataset Gallery\n"]
    
    # Add index
    content.append("## Available Datasets\n")
    for name in sorted(DATASETS.keys()):
        # Create anchor link
        anchor = name.replace("_", "-")
        content.append(f"- [{name}](#{anchor})\n")
    content.append("\n")
    
    # Add examples for each dataset
    content.append("## Examples\n")
    for name in sorted(DATASETS.keys()):
        dataset = create_dataset(name)
        
        # Add dataset header
        content.append(f"### {name}\n")
        
        # Get dataset class docstring if available
        if dataset.__class__.__doc__:
            doc = textwrap.dedent(dataset.__class__.__doc__.strip())
            content.append(f"{doc}\n")
        
        content.append("```\n")
        # Show 3 examples
        for i, item in enumerate(dataset):
            if i >= 3:
                break
            content.append(f"Example {i+1}:\n")
            content.append(f"Question: {item['question']}\n")
            content.append(f"Answer: {item['answer']}\n")
            content.append(f"Metadata: {item['metadata']}\n")
            content.append("\n")
        content.append("```\n\n")
    
    return "".join(content)


def main():
    """Generate gallery markdown file"""
    gallery_path = Path(__file__).parent.parent / "GALLERY.md"
    gallery_content = generate_gallery()
    
    with open(gallery_path, "w") as f:
        f.write(gallery_content)
    
    print(f"Generated gallery at {gallery_path}")


if __name__ == "__main__":
    main()