Performance

Dinja is designed for high performance with several optimization strategies.

Renderer Reuse

The most important performance tip: reuse Renderer instances.

# Good: Create once, reuse many times
renderer = Renderer()
for mdx_file in many_files:
    result = renderer.render(...)

# Bad: Creating new renderer for each render
for mdx_file in many_files:
    renderer = Renderer()  # Expensive!
    result = renderer.render(...)

Batch Rendering

Render multiple files in a single batch call:

# Good: Single batch call
result = renderer.render(
    Input(
        mdx={
            "file1.mdx": "...",
            "file2.mdx": "...",
            "file3.mdx": "...",
        },
        settings=Settings(output="html"),
    )
)

# Less efficient: Multiple calls
for file in files:
    result = renderer.render(
        Input(mdx={file: content}, ...)
    )

Minification

Enable minification for production:

settings = Settings(
    output="html",
    minify=True,  # Reduces output size
)

Retry Configuration

For high-load scenarios, adjust retry settings:

renderer = Renderer(
    max_retries=3,      # Default is usually sufficient
    retry_delay=0.05,   # Minimal delay
    backoff_factor=1.5  # Moderate backoff
)

Rust Performance

The Rust API provides the best performance:

// Single service instance for all renders
let service = RenderService::new(config)?;

// Batch rendering
let outcome = service.render_batch(&input)?;

Benchmarks

Dinja's Rust core provides: - 2-5x faster than creating new service instances for each render - Native performance without Node.js overhead - Efficient memory usage with renderer pooling

Best Practices

✅ Reuse Renderer instances
✅ Use batch rendering for multiple files
✅ Enable minification in production
✅ Use Rust API for maximum performance
❌ Don't create new Renderer for each render
❌ Don't render files one-by-one when batching is possible