Skip to content

This android library provide a cache with 2 layers, one in RAM in top of one disk.

License

Notifications You must be signed in to change notification settings

vincentbrison/dualcache

Repository files navigation

Deprecated

This library is now deprecated in favor of Store which fulfill all the goal of this library.

Android dualcache

API Build Status Android Arsenal

This android library provide a cache with 2 layers, one in RAM in top of one on local storage. This library is highly configurable :

Configurations Disk : Specific serializer Disk : disable
Ram : Specific serializer YES YES
Ram : References YES YES
Ram : disable YES NO
  • Specific serializer : the object stored in cache will be serialized through a serializer provided by yourself.
  • References : the objects stored in Ram are cached through there references (no serialization is done).
  • Disable : the corresponding layer (Ram or disk) is disable.

If you work with specific serializer or references you will have to provide (through an interface) the way of compute the size of cached objects, to be able to correctly execute the [LRU policy] (http://en.wikipedia.org/wiki/Cache_algorithms).

If you do not want to write your own serializer and a json serializer is enough for you, you can use dualcache-jsonserializer which will serialize object using Jackson

The following diagrams are showing how the dualcache is working :

  • DualCache with specific serializer in RAM and specific serializer in disk. dualcache-serializer

  • DualCache with references in RAM and specific serializer in disk. dualcache-serializer-ref

To get the best performance from this library, I recommend that you use larger size for the disk layer than for the Ram layer. When you try to get an object from the cache which is already in the Ram layer, the disk wont be use to keep the best performance from the Ram. If you try to get an object from the cache which is on disk and not on Ram, the object will be loaded into RAM, to ensure better further access time.

The Philosophy behind this library

When you want to use a [cache] (http://en.wikipedia.org/wiki/Cache_\(computing\)) on Android today, you have two possibilities. You whether use :

The thing is the first one only works in RAM, and the second one only on disk (internal memory of the phone). So you need to choose whether if you will use the LruCache (RAM) :

  • Very fast access to your cache.
  • High resources constraints, since the RAM allocated to your application is used for caching.
  • Not persistent among different execution of your app.

Or you will use the DiskLruCache (Disk) :

  • Slower access time than the LruCache.
  • Almost no resources constraints, since the size used on the disk (internal memory), will not impact your application.
  • Persistent among different execution of your app.

The purpose of this library is to provide both features of these two caches, by making them working together. You do not need to ask yourself anymore "Should I use this one or this one ? But this one is persistent, but the other one is faster...". With this library you only use one cache, with two layers, one in RAM, and one in Disk and you configure how they have to work to provide exactly what you need in term of caching for you application.

What's new in 3.0.0

  • Only one dependency to DiskLruCache because apk size matters.
  • Non coupled serializer, with cleaner implementation. Previously default json serializer is now a specific serializer which is available at com.vincentbrison.openlibraries.android:dualcache-jsonserializer:3.1.1.
  • Internal optimizations for better performances.
  • All the configuration is now done through Builder.
  • Better access modifiers to fully hide internal classes.

Concurrent access

Starting with version 2.2.1, the cache is supporting concurrent access. You can perform whatever operations from multiple threads and the cache takes care of the synchronization. More than that, this synchronization is optimized to block the threads only if needed, to get the best performances. In fact, put and get are synchronized on each entry, and the cache itself is locked trough a ReadWriteLock for invalidation operations.

Setup

  • Ensure you can pull artifacts from Maven Central :
repositories {
    mavenCentral()
}
  • And add to your module gradle file :
android {
    packagingOptions {
        exclude 'META-INF/LICENSE'
        exclude 'META-INF/NOTICE'
    }
}

dependencies {
    compile 'com.vincentbrison.openlibraries.android:dualcache:3.1.1'

    //compile 'com.vincentbrison.openlibraries.android:dualcache-jsonserializer:3.1.1' // If you
    // want a ready to use json serializer
}

All the configuration of the cache is done when you are building the cache through its Builder class.

Basic examples

Build your cache

First of all, you need to build you cache, through the Builder class.

  1. A cache with a serializer for RAM and disk disable :
cache = new Builder<>(CACHE_NAME, TEST_APP_VERSION, AbstractVehicule.class)
    .enableLog()
    .useSerializerInRam(RAM_MAX_SIZE, new SerializerForTesting())
    .noDisk()
    .build();
  1. A cache with references in RAM and a default serializer on disk :
cache = new Builder<>(CACHE_NAME, TEST_APP_VERSION, AbstractVehicule.class)
    .enableLog()
    .useReferenceInRam(RAM_MAX_SIZE, new SizeOfVehiculeForTesting())
    .useSerializerInDisk(DISK_MAX_SIZE, true, new DualCacheTest.SerializerForTesting(), getContext())
    .build();

You can note that when you build the cache, you need to provide an app version number. When the cache is loaded, if data exist with a inferior number, it will be invalidate. It can be extremely useful when you update your app, and change your model, to avoid crashes. This feature is possible because the DiskLruCache of Jake Wharton implemented this feature.

Put

To put an object into your cache, simply call put :

DummyClass object = new DummyClass();
object = cache.put("mykey", object);

Get

To get an object from your cache, simply call get :

DummyClass object = null;
object = cache.get("mykey");

Use cases

  • Using default serialization on RAM and on disk can be very useful for caching network exchange of data.
  • Using references in RAM and serialization on disk can be very useful to cache bitmaps.

Javadoc

The javadoc provided with this library is fully written and released on Maven.

Testing

All the configurations of the cache are (almost) fully tested through automated tests. If you fork this repo, you can launch them with the gradle command connectedAndroidTest. You need to have a device connected since the tests will be run on every device connected to your computer. An emulator or a [GenyMotion] instance is enough. A report will be available at : /{location of your fork}/lib/build/outputs/reports/androidTests/connected/index.html

License

Copyright 2016 Vincent Brison.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

   http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

About

This android library provide a cache with 2 layers, one in RAM in top of one disk.

Resources

License

Stars

Watchers

Forks

Packages

No packages published