README
¶
Cache wrapper for web applications.
The primary goal is to simplify caching of responses.
Adds guava-style loading cache and support of scopes for partial flushes.
Provides in-memory NewMemoryCache on top of hashicorp/golang-lru and
defines basic interface for other implementations.
In addition to Get and Flush methods, memory cache also support limits for a single value size, number of keys and total memory utilization. PostFlushFn adds ability to call a function on flush completion.
Install and update
go get -u github.com/go-pkgz/rest/cache
Technical details
- Cache keeps data in a simple key:value format.
- Key is a type, created with
Key(site_id)wheresite_idrepresents independent bucket in the cache. For simple cases can be set to an empty string. - Particular key set by
Key.ID(string) - Key may contain optional scopes (list of string). They not affect retrieval and used for partial (scoped) invalidation only.
- Cache is safe for concurrent use.
- Value is []byte.
Getmethod returns from the cache if the key already in. Overwise executes passed function and saves results.- Special fake implementation
cache.NopsatisfiesLoadingCacheinterface and can be used to disable any caching
Usage
// create in-memory cache with max keys=50, total (max) size=2000 and max cached size of a record = 200
lc, err := cache.NewMemoryCache(cache.MaxKeys(50), cacheMaxCacheSize(2000), cache.MaxValSize(200))
if err != nil {
panic(err)
}
...
// load cached value for key1. Call func if not cached yet or evicted
res, err = lc.Get(cache.NewKey("site1").ID("key1").Scopes("scope1"), func() ([]byte, error) {
return []byte("1234567890"), nil
})
lc.Flush(cache.Flusher("site1").Scopes("scope1")) // invalidate cache for scope1
Documentation
¶
Overview ¶
Package cache implements a wrapper on top of hashicorp/golang-lru with guava-style loader func. In addition for Get (i.e. load from cache if found and put to cache if absent) it adds a Key struct with record ID, site ID and invalidation scopes. Scopes used to evict matching records. Additional limits added for total cache size, mac number of keys and max size of the value. If exceeded it won't put to cache but will call loader func. Usually the cache involved on []byte response level, i.e. post-marshaling right before response's send.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type FlusherRequest ¶
type FlusherRequest struct {
// contains filtered or unexported fields
}
FlusherRequest used as input for cache.Flush
func Flusher ¶
func Flusher(siteID string) FlusherRequest
Flusher makes new FlusherRequest with empty scopes
func (FlusherRequest) Scopes ¶
func (f FlusherRequest) Scopes(scopes ...string) FlusherRequest
Scopes adds scopes to FlusherRequest
type Key ¶
type Key struct {
// contains filtered or unexported fields
}
Key for cache
func ParseKey ¶
ParseKey gets compound key created by Key func and split it to the actual key and scopes
type LoadingCache ¶
type LoadingCache interface {
Get(key Key, fn func() ([]byte, error)) (data []byte, err error) // load from cache if found or put to cache and return
Flush(req FlusherRequest) // evict matched records
}
LoadingCache defines interface for caching
func NewMemoryCache ¶
func NewMemoryCache(options ...Option) (LoadingCache, error)
NewMemoryCache makes memoryCache implementation
type Option ¶
type Option func(lc cacheWithOpts) error
Option func type
func MaxCacheSize ¶
MaxCacheSize functional option defines the total size of cached data. By default it is 0, which means unlimited.
func MaxKeys ¶
MaxKeys functional option defines how many keys to keep. By default it is 0, which means unlimited.
func MaxValSize ¶
MaxValSize functional option defines the largest value's size allowed to be cached By default it is 0, which means unlimited.
func PostFlushFn ¶
func PostFlushFn(postFlushFn func()) Option
PostFlushFn functional option defines how callback function called after each Flush.
Source Files