Decorator Registry¶

Overview¶

The Decorator Registry is a central component of the Prompt Decorators package that enables runtime discovery and management of decorators. It provides a mechanism to register, discover, and use decorators dynamically, making it easier to extend the system with new decorators and to find decorators by name, category, or other criteria.

Features¶

Dynamic Registration: Register decorators at runtime from code or JSON definitions
Decorator Discovery: Find decorators by name, category, or other criteria
Version Management: Support for multiple versions of the same decorator
Compatibility Checking: Verify compatibility between decorators
Category Organization: Group decorators by category for easier discovery

Implementation¶

The Decorator Registry is implemented in the prompt_decorators.utils.discovery module. The main class is DecoratorRegistry, which provides methods for registering and discovering decorators.

Key Components¶

DecoratorRegistry: The main class that manages the registry of decorators
register_decorator: Method to register a decorator class or instance
get_decorator: Method to retrieve a decorator by name
get_all_decorators: Method to retrieve all registered decorators
find_decorators_by_category: Method to find decorators by category

Usage¶

Registering Decorators¶

Decorators can be registered in several ways:

Registering a Decorator Class:

from prompt_decorators.utils.discovery import DecoratorRegistry
from prompt_decorators.decorators.generated.decorators.concise import Concise

# Get the registry instance
registry = DecoratorRegistry()

# Register a decorator class
registry.register_decorator(Concise)

Registering a Decorator Instance:

from prompt_decorators.utils.discovery import DecoratorRegistry
from prompt_decorators.decorators.generated.decorators.concise import Concise

# Get the registry instance
registry = DecoratorRegistry()

# Create a decorator instance
concise = Concise(maxWords=100, bulletPoints=True, level=2)

# Register the decorator instance
registry.register_decorator(concise)

Registering All Decorators from a Directory:

import os
import importlib
import inspect
from pathlib import Path
from prompt_decorators.utils.discovery import DecoratorRegistry
from prompt_decorators.core.base import BaseDecorator

def register_decorators():
    registry = DecoratorRegistry()
    registry.clear()

    # Path to the generated decorators directory
    decorators_dir = Path("prompt_decorators/decorators/generated/decorators")

    # Iterate through all Python files in the directory
    for file_path in decorators_dir.glob("*.py"):
        if file_path.name == "__init__.py":
            continue

        # Import the module
        module_name = f"prompt_decorators.decorators.generated.decorators.{file_path.stem}"
        try:
            module = importlib.import_module(module_name)

            # Find all classes in the module that are subclasses of BaseDecorator
            for name, obj in inspect.getmembers(module):
                if inspect.isclass(obj) and issubclass(obj, BaseDecorator) and obj != BaseDecorator:
                    # Register the decorator
                    decorator_name = getattr(obj, "name", name)
                    registry.register_decorator(obj)
                    print(f"  - Registered: {decorator_name}")
        except Exception as e:
            print(f"Error registering decorators from {module_name}: {e}")

    # Print summary
    decorators = registry.get_all_decorators()
    print(f"\nRegistered {len(decorators)} decorators:")

    # Print categories
    categories = set(decorator.category for decorator in decorators)
    print(f"\nDecorator categories ({len(categories)}):")
    for category in categories:
        category_decorators = registry.find_decorators_by_category(category)
        print(f"  - {category}: {len(category_decorators)} decorators")

Finding Decorators¶

Once decorators are registered, they can be discovered in several ways:

Getting a Decorator by Name:

from prompt_decorators.utils.discovery import DecoratorRegistry

registry = DecoratorRegistry()
concise = registry.get_decorator("Concise")

Getting All Decorators:

from prompt_decorators.utils.discovery import DecoratorRegistry

registry = DecoratorRegistry()
all_decorators = registry.get_all_decorators()

Finding Decorators by Category:

from prompt_decorators.utils.discovery import DecoratorRegistry

registry = DecoratorRegistry()
tone_decorators = registry.find_decorators_by_category("tone")

Using Registered Decorators¶

Once a decorator is retrieved from the registry, it can be used like any other decorator:

from prompt_decorators.utils.discovery import DecoratorRegistry

registry = DecoratorRegistry()
concise = registry.get_decorator("Concise")

# Create an instance of the decorator
concise_instance = concise(maxWords=100, bulletPoints=True, level=2)

# Apply the decorator to a prompt
original_prompt = "Explain the concept of quantum computing in detail."
decorated_prompt = concise_instance.apply(original_prompt)

Examples¶

For complete examples of how to use the Decorator Registry, see the following example scripts:

examples/register_all_decorators.py: Demonstrates how to register all decorators from the generated directory
examples/use_registered_decorators.py: Shows how to use registered decorators to modify prompts

Best Practices¶

Clear the Registry Before Bulk Registration: If you're registering multiple decorators at once, it's a good practice to clear the registry first to avoid duplicates.
Handle Exceptions During Registration: When registering decorators from external sources, make sure to handle exceptions properly to avoid breaking your application.
Check Decorator Existence Before Use: Always check if a decorator exists in the registry before trying to use it.
Use Categories for Organization: Organize decorators by category to make them easier to discover and use.
Consider Version Compatibility: When using multiple decorators together, consider their version compatibility to avoid unexpected behavior.

Future Enhancements¶

The Decorator Registry is designed to be extensible and can be enhanced in several ways:

Plugin System: Add support for loading decorators from external plugins
Web API: Create a web API for discovering and using decorators
Caching: Implement caching for better performance
Analytics: Add analytics to track decorator usage
UI Integration: Create a user interface for browsing and selecting decorators