0.6.5

Misc

• Updated README with examples for:
  • Using outgoing request handling.

0.6.4

Misc

• Updated README with examples for:
  • Using outgoing request handling.
• Updated pre-commit hooks versions

0.6.3

Misc

• Updated README with examples for:
  • Using context managers for async workflows.

0.6.2

Features

• Asynchronous Context Manager Support:
  • Added __aenter__ and __aexit__ methods to RatelimitIO for use as an asynchronous context manager.
  • Ensures Lua scripts are automatically loaded into Redis when entering the context.
  • Simplifies resource management and integration in async frameworks like FastAPI and aiohttp.

0.6.1

Misc

• Update version in pyproject.toml, README.md

0.6.0

Features

• Improved Error Handling:
  • Introduced specific exception classes for better error categorization:
    • RatelimitIOError: Base class for rate limit-related errors.
    • RatelimitExceededError: Raised when the rate limit is exceeded.
    • ScriptLoadError: Raised when the Lua script fails to load into Redis.
  • Enhanced error messages for clarity and debugging.
• Logging Support:
  • Added logging for critical operations and errors:
    • Logs errors when the Lua script fails to load into Redis.
    • Logs exception details for debugging unexpected issues.

Misc

• Code Simplification and Readability:
  • Refactored exception handling for better readability and maintainability.
  • Consolidated duplicated logic between synchronous and asynchronous methods.
• Documentation and Examples:
  • Updated inline comments and docstrings for clarity.
  • Included examples of error handling and logging usage.
• Testing Improvements:
  • Increased test coverage for key generation, rate limit enforcement, and exception handling.
  • Added edge case tests for LimitSpec and Redis Lua script loading.

0.5.0

Features

• Dynamic is_incoming Detection:
  • Added automatic inference for is_incoming in decorators based on usage context:
    • Incoming requests (e.g., API handlers): Raise RatelimitIOError when limits are exceeded.
    • Outgoing requests (e.g., client-side throttling): Wait until a slot is available.
  • Default behavior is now is_incoming=True, suitable for most incoming request use cases.
  • Explicit is_incoming=False can still be set during initialization for outgoing request handling.
• Removed base_url parameter:
  • Simplified the API by fully replacing base_url with default_key.
  • Unified logic for key preparation and eliminated redundancy.
• Updated _prepare_key logic:
  • Adjusted key priority to: provided_key > unique_key > default_key > kwargs["ip"] > "unknown_key".
  • Added async/sync prefixes to keys for better differentiation in mixed environments.
• Enhanced decorator functionality:
  • Supports both synchronous and asynchronous functions.
  • Automatically applies appropriate behavior for incoming or outgoing requests.

Fixes

• Fixed issues with key generation for mixed environments:
  • Resolved edge cases where default keys were incorrectly prioritized.
  • Ensured robust handling of kwargs["ip"] when no default_key or unique_key is provided.

Misc

• Updated README and examples:
  • Added detailed explanation for is_incoming behavior and dynamic detection.
  • Removed all references to base_url and clarified usage of default_key.
  • Included examples for both incoming and outgoing request scenarios.
• Improved error messages for missing rate limit specifications (limit_spec or default_limit).
• Increased test coverage to ensure robust handling of edge cases in key generation, rate limiting, and decorator functionality.

0.4.0

Fixes

• Updated default_key in RatelimitIO.__init__ to default to None instead of "unknown_key", ensuring kwargs["ip"] is used when unique_key and default_key are not provided.
• Renamed _prepare_key_and_limit to _prepare_key for clarity.
• Refactored _prepare_key to simplify logic and resolve inconsistencies in key generation.
• Fixed key handling in wait and a_wait to correctly validate and apply rate limits.
• Ensured proper rate-limiting behavior for async and sync decorators with function-specific keys.
• Addressed issues causing NoneType errors when limit_spec was not provided in wait and a_wait.

Misc

• Enhanced key generation with async/sync-specific prefixes to distinguish between function call types.
• Updated virtualenv dependency from 20.28.1 to 20.29.0 via poetry update.

0.3.1

Fixes

• Removed TypeAlias import from typing_extensions for better compatibility with Python 3.8 and 3.9, as TypeAlias is now part of typing in Python 3.10+.
• Replaced the RedisBackend alias with a direct use of Union[Redis, AsyncRedis] in the backend parameter of the __init__ method for backward compatibility.

Misc

• Updated README with improved examples for:
  • Using middleware for FastAPI: Enhanced to include proper error handling with JSONResponse for 429 responses, providing more clarity and correctness.
  • Using middleware for Django: Added a class-based middleware example leveraging MiddlewareMixin for better integration with Django's request/response lifecycle. Also demonstrated an example with DRF's APIView.

0.3.0

Features

• Introduced a custom error class RatelimitIOError, designed to mirror the behavior of HTTPException, making it easier to integrate with web frameworks like Flask, Django, and FastAPI.
• Enhanced decorator functionality
  • Automatically applies base_limit when the decorator is used without explicit arguments.
• Added seamless support for distinguishing between incoming and outgoing requests using the is_incoming flag.
• Implemented 429 Too Many Requests handling:
  • Provides examples for integrating with Flask, Django, and FastAPI to manage error responses gracefully.

Fixes

• Fixed an issue where unique_key fallback to default_key was inconsistent.
• Improved Redis script loading to ensure proper error handling and robustness during high-concurrency operations.

Misc

• Updated README with examples for:
  • Using decorators and context managers for both sync and async workflows.
  • Handling 429 Too Many Requests errors in popular frameworks.
• Enhanced test coverage to verify all added features and scenarios.