fastapi-cacher 0.3.3

A caching library for FastAPI

FastAPI-Cacher

An intuitive caching library for FastAPI inspired by Flask-Caching. It uses decorators for easy endpoint caching, supports both absolute and sliding window expiration mechanisms.

Installation

Install and update using pip

pip install -U fastapi-cacher

Cache Types

Configured in the CacheConfig class:

# Options: "SimpleCache", "RedisCache", "MongoCache", "MemCache"
cache_type = "SimpleCache"  # Default: "SimpleCache", 

SimpleCache

• Suitable for development and single-thread applications.
• Default cache type if cache_type is not specified or CacheConfig is not provided.

RedisCache

• Robust option for production environments.
• Supports distributed caching.
• Configuration requires specifying either redis_url or both redis_host and redis_password.

MongoCache

• Uses MongoDB's expireAfterSeconds index to automatically manage the expiration of cached entries.
• Requires a MongoDB connection URL.

MemCache

• Uses a Memcached server to store cached data.
• Specify memcache_host and memcache_port.

Each cache type has specific attributes in the CacheConfig that need to be configured:

from fastapi_cacher import CacheConfig

# For SimpleCache
cache_config = CacheConfig(
  cache_type="SimpleCache",
  simple_cache_threshold=100,  # Default: 100 (number of items to store in cache, before deleting the oldest)
)

# For RedisCache
cache_config = CacheConfig(
  cache_type="RedisCache",
  redis_url="redis://localhost:6379",  # either redis_url or redis_host, redis_port, redis_password
  redis_host="your_redis_host",
  redis_port=6379,  # Default: 6379
  redis_password="your_redis_password",
  redis_db=0,  # Default: 0
)

# For MongoCache
cache_config = CacheConfig(
  cache_type="MongoCache",
  mongo_url="mongodb://user:password@localhost:27017",
  mongo_database="fastapi_cache",
  mongo_collection="your_cache_collection",  # Default: "cache"
  mongo_direct_connection=False  # Default: False
)

# For MemCache
cache_config = CacheConfig(
  cache_type="MemCache",
  memcache_host="your_memcache_host",  # Default: ""
  memcache_port=11211,  # Default: 11211
  memcache_threshold=100  # Default: 100 (number of items to store in cache, before deleting the oldest)
)

Generic CacheConfig Attributes with Defaults:

• cache_type (str) = SimpleCache: Sets the caching strategy (e.g., SimpleCache, RedisCache).
• default_timeout (int) = 300: Default timeout in seconds if not specified in decorator.
• app_space (str) = fastapi-cacher: Namespace prefix for cache keys to avoid conflicts.
• coder (Coder) = JsonCoder: Serialization coder for caching.
• sliding_expiration (bool) = False: Sets the caching mechanism globally, if True, the expiration time will be reset on every access (overwritten by the decorator if specified there).
• Time Constants (ONE_HOUR, ONE_DAY, etc.): Predefined time intervals in seconds for easy setup of expiration times.

Example:

cache_config = CacheConfig(
  cache_type="SimpleCache",
  default_timeout=600,
  app_space="my_app",
  sliding_expiration=True
)

Usage

To use the caching functionality, decorate your FastAPI endpoints with the @cache.cached decorator. Here are some examples for each type of cache:

from asyncio import sleep

from fastapi import FastAPI, Request, Response
from fastapi_cacher import Cache, CacheConfig

app = FastAPI()
# Configuring with RedisCache; for settings of other cache types, see the CacheConfig section above.
cache_config = CacheConfig(
    cache_type="RedisCache",
    redis_host="your_redis_host",
    redis_password="your_redis_password",
    default_timeout=600,  # default if not specified in the decorator
    app_space="fastapi-cacher",
    sliding_expiration=False
)
cache = Cache(config=cache_config)


@app.get("/item/{item_id}")
@cache.cached(timeout=300,
              sliding_expiration=False,
              namespace="item_detail",
              query_params=True,
              json_body=False,
              require_auth_header=False)
async def get_item(request: Request, response: Response, item_id: int):
    """
    request parameter is required in the function signature for the cache to work.
    request and response parameters can be named differently:
    async def get_item(req: Request, resp: Response, item_id: int):
    """
    await sleep(3)
    return {"id": item_id, "name": "Item Name"}


@app.get("/items/")
@cache.cached(timeout=cache_config.ONE_HOUR,
              namespace="item_detail")
async def get_items(request: Request, response: Response):
    """
    request parameter is required in the function signature for the cache to work.
    request and response parameters can be named differently:
    async def get_item(req: Request, resp: Response, item_id: int):
    """
    await sleep(3)
    return {"id": 1, "name": "Item Name"}

cache.cached decorator arguments with defaults:

• timeout (int) = None: Timeout in seconds. Set to 0 to never expire. If not specified, the default timeout from the cache config is used. A pre-calculated values in the cache_config can be used, e.g., cache_config.ONE_HOUR, cache_config.ONE_DAY, etc.
• sliding_expiration (bool) = None: If True, the expiration time will be reset on every access. If set, it overwrites the cache_config setting.
• namespace (str) = "": Allows scoping of cache keys.
• query_params (book) = True: Consider URL query parameters for caching.
• json_body = False: Include requests JSON body in the cache string key.
• require_auth_header = False: Include the Authorization header in the cache string key. If set to True, the Authorization header is required in the request and if not present - Raises HTTPException(401).

More about the sliding expiration mechanism:

The sliding expiration mechanism resets the expiration time of a cached item each time it is accessed. This means the item will only be deleted if it is not accessed for the specified timeout period.

• Global Setting: Set in CacheConfig to apply by default to all endpoints.
• Individual Setting: Can be overridden in the cache.cached decorator for specific endpoints.

Clearing the Cache

Endpoints can be configured to clear the cache selectively or entirely.

@app.post('/clear-cache/')
async def clear_cache(namespace: str = None, key: str = None):
    """
    Clears the cache.

    - `namespace`: Optional. The namespace of the cache to clear.
    - `key`: Optional. The specific key within the namespace to clear.

    If no parameters are provided, the entire cache will be cleared.
    example:
    http://domain/clear-cache/?namespace=test&key=specific-key
    """
    await cache.clear(namespace=namespace, key=key)
    return "Cache cleared!"

Other cache methods:

# set 
await cache.set(key="key", value="value", timeout=300)

# get
value = await cache.get(key="key")

# get with ttl
ttl, cached_result = await cache.get_with_ttl(key="key")

Contributions

Contributions to the fastapi-cacher project are welcome. Please ensure to follow best practices for code quality and add tests for new features.

License

This project is licensed under the MIT License.

Changelog

[0.3.0] - 2024-07-10:

Added

• Support for sliding window expiration mechanism.

[0.2.0] - 2024-07-08:

Added

• Support for MongoDB cache: MongoCache.
• Support for Authorization header in cache.cached decorator.
• Support for JSON body in cache.cached decorator.
• Support for dynamic Request and Response parameters names in the function signature.

Fixed

• Issue with json parsing of MemCache results.

Changed

• default_timeout in RedisCache from 150 to 300.
• Updated README file with more examples.

[0.1.0] - 2024-06-24 initial release

Added

• Initial release of the package.