A redis plugin for cache-service.
npm install cache-service-redis* A redis plugin for cache-service
* AND a standalone redis wrapper
#### Features
* Background refresh
* Cache objects--automatic serialization/deserialization of values
* Easy config handling--pass the name of an ENV OR the redis connection config string OR all redis connection params
* Robust API
* Built-in logging with a verbose flag.
* Compatible with cache-service and superagent-cache
* Public access to the underlying node_redis instance
* A more logical API--.mset() takes an object of keys and values rather than a comma-separated argument list
* .mset() allows you to set expirations on a per key, per function call, and/or per cache-service-redis instance basis (Vanilla redis does not let .mset() set expirations at all)
* Built-in redis mock support for local testing
Require and instantiate
``javascript
var csRedis = require('cache-service-redis');
var cacheModuleConfig = {redisEnv: 'REDISCLOUD_URL'};
var redisCache = new csRedis(cacheModuleConfig);
`
Cache!
`javascript`
redisCache.set('key', 'value');
cache-service-redis's constructor takes an optional config object with any number of the following properties:
An arbitrary identifier you can assign so you know which cache is responsible for logs and errors.
* type: string
* default: 'redis'
The expiration to include when executing cache set commands. Can be overridden via .set()'s optional expiration param.
* type: int
* default: 900
* measure: seconds
This is the most generic way to pass in your redis configuraiton options.
* type: object
#### Example
`javascript`
var redisData = {
port: myRedisPort,
hostname: myRedisHostname,
auth: myRedisAuth
}
If you have all of your redis params already prepared as a URL in the following format: http://uri:password@hostname:port, then you can simply pass that URL with the object key redisUrl.
* type: string
If you have a redis URL contained in an env variable (in process.env[redisEnv]), cache-service can retrieve it for you if you pass the env variable name with the object key redisEnv.
* type: string
By default, cache-service-redis attempts to JSON.stringify and JSON.parse values on set and get respectively. If it fails, it just sets/gets the raw value. This may not always meet your needs. As such, you can provide your own object with stringify and parse public functions.
* type: object
If you want to test your cache-service-redis implementation, you can pass a redis mock with this key and cache-service-redis will consume it for testing purposes. See the Using A Redis Mock section for a more throrough explanation.
* type: object that mocks redis
How frequently should all background refresh-enabled keys be scanned to determine whether they should be refreshed. For a more thorough explanation on background refresh, see the Using Background Refresh section.
* type: int
* default: 60000
* measure: milliseconds
The maximum ttl a scanned background refresh-enabled key can have without triggering a refresh. This number should always be greater than backgroundRefreshInterval.
* type: int
* default: 70000
* measure: milliseconds
Whether to throw an exception if backgroundRefreshInterval is greater than backgroundRefreshMinTtl. Setting this property to false is highly discouraged.
* type: boolean
* default: true
> When used with cache-service, this property is overridden by cache-service's verbose value.
When false, cache-service-redis will log only errors. When true, cache-service-redis will log all activity (useful for testing and debugging).
* type: boolean
* default: false
This module automatically attempts to JSON.stringify values on set and JSON.parse values on get. In the event that either JSON function throws an error, the module assums the value being set or retrieved was not JSON and returns it as is. If you know you will only ever set and retrieve JSON and would therefore like errors to be logged when the JSON functions throw them, set this property to true.
* type: boolean
* default: false
Although this is a redis wrapper, its API differs in some small cases from redis's own API both because the redis API is sometimes dumb and because all cache-service-compatible cache modules match cache-service's API.
Retrieve a value by a given key.
* key: type: string
* callback: type: function
* err: type: object
* response: type: string or object
Retrieve the values belonging to a series of keys. If a key is not found, it will not be in response.
* keys: type: an array of strings
* callback: type: function
* err: type: object
* response: type: object, example: {key: 'value', key2: 'value2'...}
> See the Using Background Refresh section for more about the refresh and callback params.
Set a value by a given key.
* key: type: string
* callback: type: function
* expiration: type: int, measure: seconds
* refresh: type: function
* callback: type: function
Set multiple values to multiple keys
* obj: type: object, example: {'key': 'value', 'key2': 'value2', 'key3': {cacheValue: 'value3', expiration: 60}}
* callback: type: function
This function exposes a heirarchy of expiration values as follows:
* The expiration property of a key that also contains a cacheValue property will override all other expirations. (This means that, if you are caching an object, the string 'cacheValue' is a reserved property name within that object.)cacheValue
* If an object with both and expiration as properties is not present, the expiration provided to the .mset() argument list will be used.defaultExpiration
* If neither of the above is provided, each cache's will be applied.
Delete a key or an array of keys and their associated values.
* keys: type: string || array of strings
* callback: type: function
* err: type: object
* count: type: int
Flush all keys and values.
* callback: type: function
This is the underlying node_redis instance. If needed, you can access redis functions I haven't abstracted.
With a typical cache setup, you're left to find the perfect compromise between having a long expiration so that users don't have to suffer through the worst case load time, and a short expiration so data doesn't get stale. cache-service-redis eliminates the need to worry about users suffering through the longest wait time by automatically refreshing keys for you. Here's how it works:
> cache-service-redis employs an intelligent background refresh algorithm that makes it so only one dyno executes a background refresh for any given key. You should feel confident that you will not encounter multiple dynos refreshing a single key.
#### How do I turn it on?
By default, background refresh is off. It will turn itself on the first time you pass a refresh param to .set().
#### Configure
There are three options you can manipulate. See the API section for more information about them.
* backgroundRefreshInterval
*
*
#### Use
Background refresh is exposed via the .set() command as follows:
`javascript`
cacheModule.set('key', 'value', 300, refresh, cb);
If you want to pass refresh, you must also pass cb because if only four params are passed, cache-service-redis will assume the fourth param is cb.
#### The Refresh Param
###### refresh(key, cb(err, response))
* key: type: string: this is the key that is being refreshed
* cb: type: function: you must trigger this function to pass the data that should replace the current key's value
The refresh param MUST be a function that accepts key and a callback function that accepts err and response as follows:
`javascript`
var refresh = function(key, cb){
var response = goGetData();
cb(null, response);
}
You're likely to want to test your implementation of cache-service-redis. In order to write tests that will run anywhere, you'll need to use a redis mock. I recommend using redis-js. cache-service-redis natively supports mock usage as follows:
`javascript`
var redisMock = require('redis-js');
var rcModule = require('cache-service-redis');
var redisCache = new rcModule({
redisMock: redisMock,
backgroundRefreshInterval: 500
});
If you use cache-service-redis inside of AWS Lambda, ensure you follow this example so that your lambda process will terminate when complete.
`javascriptcontext
exports.handler = function (event, context, callback) {
cache.get(key, function (err, val) {
// handle value
// terminate process with
return context.succeed();
// OR terminate process with
context.callbackWaitsForEmptyEventLoop = false;
return callback(null);
});
};
If you use cache-service-redis with superagent-cache or superagent-cache-plugin inside of AWS Lambda, follow this example.
`javascriptcontext
exports.handler = function (event, context, callback) {
superagent
.get(uri)
.use(superagentCache)
.end(function (err, response){
// handle response here
// terminate process with
return context.succeed();
// OR terminate process with
context.callbackWaitsForEmptyEventLoop = false;
return callback(null);
}
);
};