Akka actors and futures: Understanding by example - java

I'm trying to learn Akka actors and futures but after reading the docs at http://akka.io
and doing http://doc.akka.io/docs/akka/2.0.2/intro/getting-started-first-java.html
I'm still having some issues with understanding. I guess calculate the value of Pi
is a thing a lot of people can relate too, but not me =). I have search around a bit
but haven't found any examples that suits me. Therefore I thought that I would take some real-life code of mine and throw it in here and exchange it for an example of how to do this with Akka.
Ok so here we go:
I have an java play2 application where I need to take some data from my DB and index it in my elasticsearch instance.
I call the DB and get the ids for the venues.
I then split the list and create a couple of callable indextasks.
After that I invoke all tasks where each task collects the venues for the assigned ids
from the db.
For each venue index it to the elasticsearch instance and make it searchable.
Done.
Application.java:
public class Application extends Controller {
private static final int VENUE_BATCH = 1000;
private static int size;
public static Result index() {
List<Long> venueIds = DbService.getAllVenueIds();
size = venueIds.size();
Logger.info("Will index " + size + " items in total.");
ExecutorService service = Executors.newFixedThreadPool(getRuntime().availableProcessors());
int startIx = 0;
Collection<Callable<Object>> indexTasks = new ArrayList<Callable<Object>>();
do {
int endIx = Math.min(startIx + VENUE_BATCH, size);
List<Long> subList = venueIds.subList(startIx, endIx);
VenueIndexTask indexTask = new VenueIndexTask(subList);
indexTasks.add(indexTask);
} while ((startIx += VENUE_BATCH) < size);
Logger.info("Invoking all tasks!");
try {
service.invokeAll(indexTasks);
} catch (InterruptedException e) {
e.printStackTrace();
}
return ok(index.render("Done indexing."));
}
}
VenueTask:
public class VenueIndexTask implements Callable<Object> {
private List<Long> idSubList;
public VenueIndexTask(List<Long> idSubList){
this.idSubList = idSubList;
Logger.debug("Creating task which will index " + idSubList.size() + " items. " +
"Range: " + rangeAsString() + ".");
}
#Override
public Object call() throws Exception {
List<Venue> venues = DbService.getVenuesForIds(idSubList);
Logger.debug("Doing some indexing: "+venues.size());
for(Venue venue : venues) {
venue.index();
}
return null;
}
private String rangeAsString() {
return "[" + idSubList.get(0) + "-" + idSubList.get(idSubList.size() - 1) + "]";
}
}
Venue:
#IndexType(name = "venue")
public class Venue extends Index {
private String name;
// Find method static for request
public static Finder<Venue> find = new Finder<Venue>(Venue.class);
public Venue() {
}
public Venue(String id, String name) {
super.id = id;
this.name = name;
}
#Override
public Map toIndex() {
HashMap map = new HashMap();
map.put("id", super.id);
map.put("name", name);
return map;
}
#Override
public Indexable fromIndex(Map map) {
if (map == null) {
return this;
}
this.name = (String) map.get("name");
return this;
}
}
So all you Akka people out there go nuts! And please do as much as you can, propose cool futures functionality that could be used or any other knowledge/code that I could use to learn this stuff.

How I like to think of Akka (or any other message based systems) is to think like a conveyor belt, like in factories. A simplified way of thinking in Actors could be taking a pizza Order.
You, the hungry customer (Actor/Role) sends a order (A Message) to Pizza Shop
Customer service (Actor/Role) takes your order, gives you the order number (Future)
If you were impatient, you might've waited on the phone/internet/shop till you got your pizza (A synchronous/blocking transaction) otherwise you would be happy with the order number an check up on it later (non-blocking)
Customer service sends the message to the Chefs (Actor) under the supervision of Kitchen Manager (Actor). This is a very process heavy kitchen, with hierarchy. Akka likes that. See Supervision
Chef creates a new Pizza and attaches the details of the order (A new message) and passes that to the delivery boy (Actor) via the delivery manager (Supervisor Actor).
During this process, your order details haven't changed, that would be a nightmare. You would not be happy if you had pepperoni if you wanted plain cheese! All messages should be immutable! However, it may be that the message might be different for different actors. A delivery boy would expect a Pizza and the order details attached, the chef would expect an order. When a message needs to change, a new message is created.
Each actor is good at one role, how effective would it be if one guy had to do all the tasks? It may be that some actors out number others (e.q. 10 threads for Chefs, 2 for Delivery boys, 1 Customer Service).
Blocking behavior is a pain, imagine customer service is waiting for the chef and delivery boy before seeing the next customer?
Hopefully I've helped you a little, this is a huge topic and large change of mind. Good luck

Coursera currently runs a course on reactive programming which has the 3 last lectures on Akka and the actor model. This includes video lectures and homework (in Scala though not Java).
While you are too late to receive the full certificate, you can still join the course and just check the last three weeks.
https://class.coursera.org/reactive-001/class

Related

Tracking the progress between Queues in a Map

I have currently two queues and items traveling between them. Initially, an item gets put into firstQueue, then one of three dedicated thread moves it to secondQueue and finally another dedicated thread removes it. These moves obviously include some processing. I need to be able to get the status of any item (IN_FIRST, AFTER_FIRST, IN_SECOND, AFTER_SECOND, or ABSENT) and I implemented it manually by doing the update of the statusMap where the queue gets modified like
while (true) {
Item i = firstQueue.take();
statusMap.put(i, AFTER_FIRST);
process(i);
secondQueue.add(i);
statusMap.put(i, IN_SECOND);
}
This works, but it's ugly and leaves a time window where the status is inconsistent. The inconsistency is no big deal and it'd solvable by synchronization, but this could backfire as the queue is of limited capacity and may block. The ugliness bothers me more.
Efficiency hardly matters as the processing takes seconds. Dedicated threads are used in order to control concurrency. No item should ever be in multiple states (but this is not very important and not guaranteed by my current racy approach). There'll be more queues (and states) and they'll of different kinds (DelayQueue, ArrayBlockingQueue, and maybe PriorityQueue).
I wonder if there's a nice solution generalizable to multiple queues?
Does it make sense to wrap the queues with logic to manage the Item status?
public class QueueWrapper<E> implements BlockingQueue<E> {
private Queue<E> myQueue = new LinkedBlockingQueue<>();
private Map<E, Status> statusMap;
public QueueWrapper(Map<E, Status> statusMap) {
this.statusMap = statusMap;
}
[...]
#Override
public E take() throws InterruptedException {
E result = myQueue.take();
statusMap.put(result, Status.AFTER_FIRST);
return result;
}
That way status management is always related to (and contained in) queue operations...
Obviously statusMap needs to be synchronized, but that would be an issue anyway.
I see that your model might be improved in consistency, state control, and scaling.
A way of to implement this is accouple the item to your state, enqueue and dequeue this couple and create a mechanism to ensure state change.
My proposal can be see in figure below:
According with this model and your example, we can to do:
package stackoverflow;
import java.util.concurrent.LinkedBlockingQueue;
import stackoverflow.item.ItemState;
import stackoverflow.task.CreatingTask;
import stackoverflow.task.FirstMovingTask;
import stackoverflow.task.SecondMovingTask;
public class Main {
private static void startTask(String name, Runnable r){
Thread t = new Thread(r, name);
t.start();
}
public static void main(String[] args) {
//create queues
LinkedBlockingQueue<ItemState> firstQueue = new LinkedBlockingQueue<ItemState>();
LinkedBlockingQueue<ItemState> secondQueue = new LinkedBlockingQueue<ItemState>();
//start three threads
startTask("Thread#1", new CreatingTask(firstQueue));
startTask("Thread#2", new FirstMovingTask(firstQueue, secondQueue));
startTask("Thread#3", new SecondMovingTask(secondQueue));
}
}
Each task runs the operations op() of according with below affirmation on ItemState:
one of three dedicated thread moves it to secondQueue and finally
another dedicated thread removes it.
ItemState is a immutable object that contains Item and your State. This ensures consistency between Item and State values.
ItemState has acknowledgement about the next state creating a mechanism of self-controled state:
public class FirstMovingTask {
//others codes
protected void op() {
try {
//dequeue
ItemState is0 = new ItemState(firstQueue.take());
System.out.println("Item " + is0.getItem().getValue() + ": " + is0.getState().getValue());
//process here
//enqueue
ItemState is1 = new ItemState(is0);
secondQueue.add(is1);
System.out.println("Item " + is1.getItem().getValue() + ": " + is1.getState().getValue());
} catch (InterruptedException e) {
e.printStackTrace();
}
}
//others codes
}
With ItemState implemetation:
public class ItemStateImpl implements ItemState {
private final Item item;
private final State state;
public ItemStateImpl(Item i){
this.item = i;
this.state = new State();
}
public ItemStateImpl(ItemState is) {
this.item = is.getItem();
this.state = is.getState().next();
}
// gets attrs
}
So this way is possible build solutions more elegant, flexible and scalable.
Scalable because you can to control more states only changing next() and generalizing the moving task for increase the number of queue.
Results:
Item 0: AFTER_FIRST
Item 0: IN_FIRST
Item 0: IN_SECOND
Item 0: AFTER_SECOND
Item 1: IN_FIRST
Item 1: AFTER_FIRST
Item 1: IN_SECOND
Item 1: AFTER_SECOND
Item 2: IN_FIRST
Item 2: AFTER_FIRST
Item 2: IN_SECOND
... others
UPDATE(06/07/2018): analysing the use of map for search
Search in map using equals values like comparator might not work because usally the mapping between values and identity (key/hash) is not one-to-one(see figure bellow). In this way is need to create an sorted list for search values which results in O(n) (worst-case).
with Item.getValuesHashCode():
private int getValuesHashCode(){
return new HashCodeBuilder().append(value).hashCode();
}
In this case, you must keep Vector<ItemState> instead of Item and to use the key like the result of getValuesHashCode. Change the mechanism of state-control for keep first reference of the Item and the state current. See bellow:
//Main.class
public static void main(String[] args) {
... others code ...
//references repository
ConcurrentHashMap<Integer, Vector<ItemState>> statesMap = new ConcurrentHashMap<Integer, Vector<ItemState>>();
//start three threads
startTask("Thread#1", new CreatingTask(firstQueue, statesMap));
... others code ...
}
//CreateTask.class
protected void op() throws InterruptedException {
//create item
ItemState is = new ItemStateImpl(new Item(i++, NameGenerator.name()));
//put in monitor and enqueue
int key = is.getHashValue();
Vector<ItemState> items = map.get(key);
if (items == null){
items = new Vector<>();
map.put(key, items);
}
items.add(is);
//enqueue
queue.put(is);
}
//FirstMovingTask.class
protected void op() throws InterruptedException{
//dequeue
ItemState is0 = firstQueue.take();
//process
ItemState is1 = process(is0.next());
//enqueue
secondQueue.put(is1.next());
}
//ItemState.class
public ItemState next() {
//required for consistent change state
synchronized (state) {
state = state.next();
return this;
}
}
To search you must use concurrentMapRef.get(key). The result will the reference of updated ItemState.
Results in my tests for :
# key = hash("a")
# concurrentMapRef.get(key)
...
Item#7#0 : a - IN_FIRST
... many others lines
Item#7#0 : a - AFTER_FIRST
Item#12#1 : a - IN_FIRST
... many others lines
Item#7#0 : a - IN_SECOND
Item#12#1 : a - IN_FIRST
... many others lines
Item#7#0 : a - AFTER_SECOND
Item#12#1 : a - IN_FIRST
More details in code: https://github.com/ag-studies/stackoverflow-queue
UPDATED IN 06/09/2018: redesign
Generalizing this project, I can undestand that the state machine is something like:
In this way I decoupled the workers of the queues for improve concepts. I used an MemoryRep for keep the unique reference for item in overall processment.
Of course that you can use strategies event-based if you need keep ItemState in a physic repository.
This keep the previous idea and creates more legibility for the concepts. See this:
I understand that each job will have two queue (input/output) and relationship with a business model! The researcher will always find the most updated and consistent state of Item.
So, answering your ask:
I can find the consistent state of Item anywhere using MemoryRep (basically an Map), wrapping state and item in ItemState, and controlling the change state on job on enqueue or dequeue it.
The performace is keeped, except on running of next()
The state is allways consistent (for your problem)
In this model is possible use any queue type, any number of jobs/queues, and any number of state.
Additionaly this is beautiful!!
As previously answered, Wrap the queues or the item would be viable solutions or both.
public class ItemWrapper<E> {
E item;
Status status;
public ItemWrapper(Item i, Status s){ ... }
public setStatus(Status s){ ... }
// not necessary if you use a queue wrapper (see queue wrapper)
public boolean equals(Object obj) {
if ( obj instanceof ItemWrapper)
return item.equals(((ItemWrapper) obj).item)
return false;
}
public int hashCode(){
return item;
}
}
...
process(item) // process update status in the item
...
Probably a better way, already answered, is to have a QueueWrapper who update the queue status. For the fun I don't use a status map but I use the previously itemwrapper it seems cleaner (a status map works too).
public class QueueWrapper<E> implements Queue<E> {
private Queue<ItemWrapper<E>> myQueue;
static private Status inStatus; // FIRST
static private Status outStatus; // AFTER_FIRST
public QueueWrapper(Queue<E> myQueue, Status inStatus, Status outStatus) {...}
#Override
public boolean add(E e) {
return myQueue.add(new ItemWrapper(e, inStatus));
}
#Override
public E remove(){
ItemWrapper<E> result = myQueue.remove();
result.setStatus(outStatus)
return result.item;
}
...
}
You can also use AOP to inject status update in your queues without changing your queues (a status map should be more appropriate than itemwrapper).
Maybe I didn't answer well your question because an easy way to know where is your item could be to check in each queue with "contains" function.
Here's something different from what others have said. Taking from the world of queue services and systems we have the concept of message acknowledgement. This is nice, because it also gives you some built in retry logic.
I'll lay out how it would work from a high level, and if you need I can add code.
Essentially you'll have a Set to go with each of your queues. You'll wrap your queues in an object so that when you dequeue an item a few things happen
The item is removed from the queue
The item is added to the associated set
A task (lambda containing an atomic boolean (default false)) is scheduled. When run it will remove item from the set and if the boolean is false, put it back in the queue
The item and a wrapper around the boolean are returned to the caller
Once process(i); completes, your code will indicate receipt acknowledgement to the wrapper, and the wrapper will remove the item from the set and make the boolean false.
A method to return status would simply check which queue or set the item is in.
Note that this gives "at least once" delivery, meaning an item will be processed at least once, but potentially more than once if the processing time is too close to the timeout.

Complex custom Collector with Java 8

I have a stream of objects which I would like to collect the following way.
Let's say we are handling forum posts:
class Post {
private Date time;
private Data data
}
I want to create a list which groups posts by a period. If there were no posts for X minutes, create a new group.
class PostsGroup{
List<Post> posts = new ArrayList<> ();
}
I want to get a List<PostGroups> containing the posts grouped by the interval.
Example: interval of 10 minutes.
Posts:
[{time:x, data:{}}, {time:x + 3, data:{}} , {time:x + 12, data:{}, {time:x + 45, data:{}}}]
I want to get a list of posts group:
[
{posts : [{time:x, data:{}}, {time:x + 3, data:{}}, {time:x + 12, data:{}]]},
{posts : [{time:x + 45, data:{}]}
]
notice that the first group lasted till X + 22. Then a new post was received at X + 45.
Is this possible?
This problem could be easily solved using the groupRuns method of my StreamEx library:
long MAX_INTERVAL = TimeUnit.MINUTES.toMillis(10);
StreamEx.of(posts)
.groupRuns((p1, p2) -> p2.time.getTime() - p1.time.getTime() <= MAX_INTERVAL)
.map(PostsGroup::new)
.toList();
I assume that you have a constructor
class PostsGroup {
private List<Post> posts;
public PostsGroup(List<Post> posts) {
this.posts = posts;
}
}
The StreamEx.groupRuns method takes a BiPredicate which is applied to two adjacent input elements and returns true if they must be grouped together. This method creates the stream of lists where each list represents the group. This method is lazy and works fine with parallel streams.
You need to retain state between stream entries and write yourself a grouping classifier. Something like this would be a good start.
class Post {
private final long time;
private final String data;
public Post(long time, String data) {
this.time = time;
this.data = data;
}
#Override
public String toString() {
return "Post{" + "time=" + time + ", data=" + data + '}';
}
}
public void test() {
System.out.println("Hello");
long t = 0;
List<Post> posts = Arrays.asList(
new Post(t, "One"),
new Post(t + 1000, "Two"),
new Post(t + 10000, "Three")
);
// Group every 5 seconds.
Map<Long, List<Post>> gouped = posts
.stream()
.collect(Collectors.groupingBy(new ClassifyByTimeBetween(5000)));
gouped.entrySet().stream().forEach((e) -> {
System.out.println(e.getKey() + " -> " + e.getValue());
});
}
class ClassifyByTimeBetween implements Function<Post, Long> {
final long delay;
long currentGroupBy = -1;
long lastDateSeen = -1;
public ClassifyByTimeBetween(long delay) {
this.delay = delay;
}
#Override
public Long apply(Post p) {
if (lastDateSeen >= 0) {
if (p.time > lastDateSeen + delay) {
// Grab this one.
currentGroupBy = p.time;
}
} else {
// First time - start there.
currentGroupBy = p.time;
}
lastDateSeen = p.time;
return currentGroupBy;
}
}
Since no one has provided a solution with a custom collector as it was required in the original problem statement, here is a collector-implementation that groups Post objects based on the provided time-interval.
Date class mentioned in the question is obsolete since Java 8 and not recommended to be used in new projects. Hence, LocalDateTime will be utilized instead.
Post & PostGroup
For testing purposes, I've used Post implemented as a Java 16 record (if you substitute it with a class, the overall solution will be fully compliant with Java 8):
public record Post(LocalDateTime dateTime) {}
Also, I've enhanced the PostGroup object. My idea is that it should be capable to decide whether the offered Post should be added to the list of posts or rejected as the Information expert principle suggests (in short: all manipulations with the data should happen only inside a class to which that data belongs).
To facilitate this functionality, two extra fields were added: interval of type Duration from the java.time package to represent the maximum interval between the earliest post and the latest post in a group, and intervalBound of type LocalDateTime which gets initialized after the first post will be added a later on will be used internally by the method isWithinInterval() to check whether the offered post fits into the interval.
public class PostsGroup {
private Duration interval;
private LocalDateTime intervalBound;
private List<Post> posts = new ArrayList<>();
public PostsGroup(Duration interval) {
this.interval = interval;
}
public boolean tryAdd(Post post) {
if (posts.isEmpty()) {
intervalBound = post.dateTime().plus(interval);
return posts.add(post);
} else if (isWithinInterval(post)) {
return posts.add(post);
}
return false;
}
public boolean isWithinInterval(Post post) {
return post.dateTime().isBefore(intervalBound);
}
#Override
public String toString() {
return "PostsGroup{" + posts + '}';
}
}
I'm making two assumptions:
All posts in the source are sorted by time (if it is not the case, you should introduce sorted() operation in the pipeline before collecting the results);
Posts need to be collected into the minimum number of groups, as a consequence of this it's not possible to split this task and execute stream in parallel.
Building a Custom Collector
We can create a custom collector either inline by using one of the versions of the static method Collector.of() or by defining a class that implements the Collector interface.
These parameters have to be provided while creating a custom collector:
Supplier Supplier<A> is meant to provide a mutable container which store elements of the stream. In this case, ArrayDeque (as an implementation of the Deque interface) will be handy as a container to facilitate the convenient access to the most recently added element, i.e. the latest PostGroup.
Accumulator BiConsumer<A,T> defines how to add elements into the container provided by the supplier. For this task, we need to provide the logic on that will allow determining whether the next element from the stream (i.e. the next Post) should go into the last PostGroup in the Deque, or a new PostGroup needs to be allocated for it.
Combiner BinaryOperator<A> combiner() establishes a rule on how to merge two containers obtained while executing stream in parallel. Since this operation is treated as not parallelizable, the combiner is implemented to throw an AssertionError in case of parallel execution.
Finisher Function<A,R> is meant to produce the final result by transforming the mutable container. The finisher function in the code below turns the container, a deque containing the result, into an immutable list.
Note: Java 16 method toList() is used inside the finisher function, for Java 8 it can be replaced with collect(Collectors.toUnmodifiableList()) or collect(Collectors.toList()).
Characteristics allow providing additional information, for instance Collector.Characteristics.UNORDERED which is used in this case denotes that the order in which partial results of the reduction produced while executing in parallel is not significant. In this case, collector doesn't require any characteristics.
The method below is responsible for generating the collector based on the provided interval.
public static Collector<Post, ?, List<PostsGroup>> groupPostsByInterval(Duration interval) {
return Collector.of(
ArrayDeque::new,
(Deque<PostsGroup> deque, Post post) -> {
if (deque.isEmpty() || !deque.getLast().tryAdd(post)) { // if no groups have been created yet or if adding the post into the most recent group fails
PostsGroup postsGroup = new PostsGroup(interval);
postsGroup.tryAdd(post);
deque.addLast(postsGroup);
}
},
(Deque<PostsGroup> left, Deque<PostsGroup> right) -> { throw new AssertionError("should not be used in parallel"); },
(Deque<PostsGroup> deque) -> deque.stream().collect(Collectors.collectingAndThen(Collectors.toUnmodifiableList())));
}
main() - demo
public static void main(String[] args) {
List<Post> posts =
List.of(new Post(LocalDateTime.of(2022,4,28,15,0)),
new Post(LocalDateTime.of(2022,4,28,15,3)),
new Post(LocalDateTime.of(2022,4,28,15,5)),
new Post(LocalDateTime.of(2022,4,28,15,8)),
new Post(LocalDateTime.of(2022,4,28,15,12)),
new Post(LocalDateTime.of(2022,4,28,15,15)),
new Post(LocalDateTime.of(2022,4,28,15,18)),
new Post(LocalDateTime.of(2022,4,28,15,27)),
new Post(LocalDateTime.of(2022,4,28,15,48)),
new Post(LocalDateTime.of(2022,4,28,15,54)));
Duration interval = Duration.ofMinutes(10);
List<PostsGroup> postsGroups = posts.stream()
.collect(groupPostsByInterval(interval));
postsGroups.forEach(System.out::println);
}
Output:
PostsGroup{[Post[dateTime=2022-04-28T15:00], Post[dateTime=2022-04-28T15:03], Post[dateTime=2022-04-28T15:05], Post[dateTime=2022-04-28T15:08]]}
PostsGroup{[Post[dateTime=2022-04-28T15:12], Post[dateTime=2022-04-28T15:15], Post[dateTime=2022-04-28T15:18]]}
PostsGroup{[Post[dateTime=2022-04-28T15:27]]}
PostsGroup{[Post[dateTime=2022-04-28T15:48], Post[dateTime=2022-04-28T15:54]]}
You can also play around with this Online Demo

Object-Oriented design simulating a process with states [duplicate]

I have something to do for work and I need your help.
We want to implement a FSM - Finite State Machine, to identify char sequence(like: A, B, C, A, C), and tell if it accepted.
We think to implement three classes: State, Event and Machine.
The state class presents a node in the FSM, we thought to implement it with State design pattern, every node will extend from the abstract class state and every class would handle different types of events and indicate transitions to a new state. Is it good idea in your opinion?
Second thing, we don't know how to save all the transitions. Again we thought to implement it with some kind of map, that hold the starting point and gets some kind of vector with the next states, but I'm not sure thats a good idea.
I would be happy to get some ideas of how to implement it or maybe you can give me some starting points.
How should I save the FSM, meaning how should I build the tree at the beginning of the program?
I googled it and found a lot of examples but nothing that helps me.
Thanks a lot.
The heart of a state machine is the transition table, which takes a state and a symbol (what you're calling an event) to a new state. That's just a two-index array of states. For sanity and type safety, declare the states and symbols as enumerations. I always add a "length" member in some way (language-specific) for checking array bounds. When I've hand-coded FSM's, I format the code in row and column format with whitespace fiddling. The other elements of a state machine are the initial state and the set of accepting states. The most direct implementation of the set of accepting states is an array of booleans indexed by the states. In Java, however, enumerations are classes, and you can specify an argument "accepting" in the declaration for each enumerated value and initialize it in the constructor for the enumeration.
For the machine type, you can write it as a generic class. It would take two type arguments, one for the states and one for the symbols, an array argument for the transition table, a single state for the initial. The only other detail (though it's critical) is that you have to call Enum.ordinal() to get an integer suitable for indexing the transition array, since you there's no syntax for directly declaring an array with a enumeration index (though there ought to be).
To preempt one issue, EnumMap won't work for the transition table, because the key required is a pair of enumeration values, not a single one.
enum State {
Initial( false ),
Final( true ),
Error( false );
static public final Integer length = 1 + Error.ordinal();
final boolean accepting;
State( boolean accepting ) {
this.accepting = accepting;
}
}
enum Symbol {
A, B, C;
static public final Integer length = 1 + C.ordinal();
}
State transition[][] = {
// A B C
{
State.Initial, State.Final, State.Error
}, {
State.Final, State.Initial, State.Error
}
};
EasyFSM is a dynamic Java Library which can be used to implement an FSM.
You can find documentation for the same at :
Finite State Machine in Java
Also, you can download the library at :
Java FSM Library : DynamicEasyFSM
You can implement Finite State Machine in two different ways.
Option 1:
Finite State machine with a pre-defined workflow : Recommended if you know all states in advance and state machine is almost fixed without any changes in future
Identify all possible states in your application
Identify all the events in your application
Identify all the conditions in your application, which may lead state transition
Occurrence of an event may cause transitions of state
Build a finite state machine by deciding a workflow of states & transitions.
e.g If an event 1 occurs at State 1, the state will be updated and machine state may still be in state 1.
If an event 2 occurs at State 1, on some condition evaluation, the system will move from State 1 to State 2
This design is based on State and Context patterns.
Have a look at Finite State Machine prototype classes.
Option 2:
Behavioural trees: Recommended if there are frequent changes to state machine workflow. You can dynamically add new behaviour without breaking the tree.
The base Task class provides a interface for all these tasks, the leaf tasks are the ones just mentioned, and the parent tasks are the interior nodes that decide which task to execute next.
The Tasks have only the logic they need to actually do what is required of them, all the decision logic of whether a task has started or not, if it needs to update, if it has finished with success, etc. is grouped in the TaskController class, and added by composition.
The decorators are tasks that “decorate” another class by wrapping over it and giving it additional logic.
Finally, the Blackboard class is a class owned by the parent AI that every task has a reference to. It works as a knowledge database for all the leaf tasks
Have a look at this article by Jaime Barrachina Verdia for more details
Hmm, I would suggest that you use Flyweight to implement the states. Purpose: Avoid the memory overhead of a large number of small objects. State machines can get very, very big.
http://en.wikipedia.org/wiki/Flyweight_pattern
I'm not sure that I see the need to use design pattern State to implement the nodes. The nodes in a state machine are stateless. They just match the current input symbol to the available transitions from the current state. That is, unless I have entirely forgotten how they work (which is a definite possiblilty).
If I were coding it, I would do something like this:
interface FsmNode {
public boolean canConsume(Symbol sym);
public FsmNode consume(Symbol sym);
// Other methods here to identify the state we are in
}
List<Symbol> input = getSymbols();
FsmNode current = getStartState();
for (final Symbol sym : input) {
if (!current.canConsume(sym)) {
throw new RuntimeException("FSM node " + current + " can't consume symbol " + sym);
}
current = current.consume(sym);
}
System.out.println("FSM consumed all input, end state is " + current);
What would Flyweight do in this case? Well, underneath the FsmNode there would probably be something like this:
Map<Integer, Map<Symbol, Integer>> fsm; // A state is an Integer, the transitions are from symbol to state number
FsmState makeState(int stateNum) {
return new FsmState() {
public FsmState consume(final Symbol sym) {
final Map<Symbol, Integer> transitions = fsm.get(stateNum);
if (transisions == null) {
throw new RuntimeException("Illegal state number " + stateNum);
}
final Integer nextState = transitions.get(sym); // May be null if no transition
return nextState;
}
public boolean canConsume(final Symbol sym) {
return consume(sym) != null;
}
}
}
This creates the State objects on a need-to-use basis, It allows you to use a much more efficient underlying mechanism to store the actual state machine. The one I use here (Map(Integer, Map(Symbol, Integer))) is not particulary efficient.
Note that the Wikipedia page focuses on the cases where many somewhat similar objects share the similar data, as is the case in the String implementation in Java. In my opinion, Flyweight is a tad more general, and covers any on-demand creation of objects with a short life span (use more CPU to save on a more efficient underlying data structure).
Consider the easy, lightweight Java library EasyFlow. From their docs:
With EasyFlow you can:
implement complex logic but keep your code simple and clean
handle asynchronous calls with ease and elegance
avoid concurrency by using event-driven programming approach
avoid StackOverflow error by avoiding recursion
simplify design, programming and testing of complex java applications
I design & implemented a simple finite state machine example with java.
IFiniteStateMachine: The public interface to manage the finite state machine
such as add new states to the finite state machine or transit to next states by specific actions.
interface IFiniteStateMachine {
void setStartState(IState startState);
void setEndState(IState endState);
void addState(IState startState, IState newState, Action action);
void removeState(String targetStateDesc);
IState getCurrentState();
IState getStartState();
IState getEndState();
void transit(Action action);
}
IState: The public interface to get state related info
such as state name and mappings to connected states.
interface IState {
// Returns the mapping for which one action will lead to another state
Map<String, IState> getAdjacentStates();
String getStateDesc();
void addTransit(Action action, IState nextState);
void removeTransit(String targetStateDesc);
}
Action: the class which will cause the transition of states.
public class Action {
private String mActionName;
public Action(String actionName) {
mActionName = actionName;
}
String getActionName() {
return mActionName;
}
#Override
public String toString() {
return mActionName;
}
}
StateImpl: the implementation of IState. I applied data structure such as HashMap to keep Action-State mappings.
public class StateImpl implements IState {
private HashMap<String, IState> mMapping = new HashMap<>();
private String mStateName;
public StateImpl(String stateName) {
mStateName = stateName;
}
#Override
public Map<String, IState> getAdjacentStates() {
return mMapping;
}
#Override
public String getStateDesc() {
return mStateName;
}
#Override
public void addTransit(Action action, IState state) {
mMapping.put(action.toString(), state);
}
#Override
public void removeTransit(String targetStateDesc) {
// get action which directs to target state
String targetAction = null;
for (Map.Entry<String, IState> entry : mMapping.entrySet()) {
IState state = entry.getValue();
if (state.getStateDesc().equals(targetStateDesc)) {
targetAction = entry.getKey();
}
}
mMapping.remove(targetAction);
}
}
FiniteStateMachineImpl: Implementation of IFiniteStateMachine. I use ArrayList to keep all the states.
public class FiniteStateMachineImpl implements IFiniteStateMachine {
private IState mStartState;
private IState mEndState;
private IState mCurrentState;
private ArrayList<IState> mAllStates = new ArrayList<>();
private HashMap<String, ArrayList<IState>> mMapForAllStates = new HashMap<>();
public FiniteStateMachineImpl(){}
#Override
public void setStartState(IState startState) {
mStartState = startState;
mCurrentState = startState;
mAllStates.add(startState);
// todo: might have some value
mMapForAllStates.put(startState.getStateDesc(), new ArrayList<IState>());
}
#Override
public void setEndState(IState endState) {
mEndState = endState;
mAllStates.add(endState);
mMapForAllStates.put(endState.getStateDesc(), new ArrayList<IState>());
}
#Override
public void addState(IState startState, IState newState, Action action) {
// validate startState, newState and action
// update mapping in finite state machine
mAllStates.add(newState);
final String startStateDesc = startState.getStateDesc();
final String newStateDesc = newState.getStateDesc();
mMapForAllStates.put(newStateDesc, new ArrayList<IState>());
ArrayList<IState> adjacentStateList = null;
if (mMapForAllStates.containsKey(startStateDesc)) {
adjacentStateList = mMapForAllStates.get(startStateDesc);
adjacentStateList.add(newState);
} else {
mAllStates.add(startState);
adjacentStateList = new ArrayList<>();
adjacentStateList.add(newState);
}
mMapForAllStates.put(startStateDesc, adjacentStateList);
// update mapping in startState
for (IState state : mAllStates) {
boolean isStartState = state.getStateDesc().equals(startState.getStateDesc());
if (isStartState) {
startState.addTransit(action, newState);
}
}
}
#Override
public void removeState(String targetStateDesc) {
// validate state
if (!mMapForAllStates.containsKey(targetStateDesc)) {
throw new RuntimeException("Don't have state: " + targetStateDesc);
} else {
// remove from mapping
mMapForAllStates.remove(targetStateDesc);
}
// update all state
IState targetState = null;
for (IState state : mAllStates) {
if (state.getStateDesc().equals(targetStateDesc)) {
targetState = state;
} else {
state.removeTransit(targetStateDesc);
}
}
mAllStates.remove(targetState);
}
#Override
public IState getCurrentState() {
return mCurrentState;
}
#Override
public void transit(Action action) {
if (mCurrentState == null) {
throw new RuntimeException("Please setup start state");
}
Map<String, IState> localMapping = mCurrentState.getAdjacentStates();
if (localMapping.containsKey(action.toString())) {
mCurrentState = localMapping.get(action.toString());
} else {
throw new RuntimeException("No action start from current state");
}
}
#Override
public IState getStartState() {
return mStartState;
}
#Override
public IState getEndState() {
return mEndState;
}
}
example:
public class example {
public static void main(String[] args) {
System.out.println("Finite state machine!!!");
IState startState = new StateImpl("start");
IState endState = new StateImpl("end");
IFiniteStateMachine fsm = new FiniteStateMachineImpl();
fsm.setStartState(startState);
fsm.setEndState(endState);
IState middle1 = new StateImpl("middle1");
middle1.addTransit(new Action("path1"), endState);
fsm.addState(startState, middle1, new Action("path1"));
System.out.println(fsm.getCurrentState().getStateDesc());
fsm.transit(new Action(("path1")));
System.out.println(fsm.getCurrentState().getStateDesc());
fsm.addState(middle1, endState, new Action("path1-end"));
fsm.transit(new Action(("path1-end")));
System.out.println(fsm.getCurrentState().getStateDesc());
fsm.addState(endState, middle1, new Action("path1-end"));
}
}
Full example on Github
Well this is an old question but while nobody mentioned here, I will advice to check two existing frameworks before you implement you own State Machines.
One is Spring State Machine most of you are familiar with Spring framework, which allow us to use several features of Spring like dependency injection and everything else that Spring can offer.
It is really great for modelling the lifecycle of an Apparat, with states like INITIALIZING, STARTED, ERROR, RECOVERING, SHUTTINGDOWN, etc.. but I see lots of people are trying to model a Shopping Chart, a Reservation System with it, the memory footprint a Spring State Machine is relatively big to model millions of Shopping Charts or Reservations.
One other disadvantage, Spring State Machine, while has a capability to persist itself for long running processes but it does not have any mechanism to adapt to changes in these processes, if you persist a process and you have to recover it lets say 10 days later with a change occurred in your business process because of new software release / requirement, you have no built in means to deal with it.
I have several blogs, blog1 blog2, demonstrating how you can program Spring State Machine, specially model driven way, if you want to check it.
Mainly because the disadvantages I mentioned, I advice you to look another framework first, Akka FSM (Finite State Machine) which is more fitting with its low memory footprint to have millions and millions of instances and has a capability to adapt changing long running processes.
Now you can develop with Akka framework with Java but believe me because of some missing language elements, you don't want to read the produced code, Scala is a much more fitting language to develop with Akka. Now I hear you saying Scala is too complex, I can't convince my project leads to develop with Scala, to convince you all this is an option, I developed a Proof of Concept application using a Java/Scala hybrid with all Scala Akka Finite State Machine code generated from an UML model, if you want to check it out here the links to the blogs, blog3 blog4.
I hope this information would help you.
Here is a SUPER SIMPLE implementation/example of a FSM using just "if-else"s which avoids all of the above subclassing answers (taken from Using Finite State Machines for Pattern Matching in Java, where he is looking for a string which ends with "#" followed by numbers followed by "#"--see state graph here):
public static void main(String[] args) {
String s = "A1#312#";
String digits = "0123456789";
int state = 0;
for (int ind = 0; ind < s.length(); ind++) {
if (state == 0) {
if (s.charAt(ind) == '#')
state = 1;
} else {
boolean isNumber = digits.indexOf(s.charAt(ind)) != -1;
if (state == 1) {
if (isNumber)
state = 2;
else if (s.charAt(ind) == '#')
state = 1;
else
state = 0;
} else if (state == 2) {
if (s.charAt(ind) == '#') {
state = 3;
} else if (isNumber) {
state = 2;
} else if (s.charAt(ind) == '#')
state = 1;
else
state = 0;
} else if (state == 3) {
if (s.charAt(ind) == '#')
state = 1;
else
state = 0;
}
}
} //end for loop
if (state == 3)
System.out.println("It matches");
else
System.out.println("It does not match");
}
P.S: Does not answer your question directly, but shows you how to implement a FSM very easily in Java.

Making massive amounts of individual row updates faster or more efficient

I'm writing a java application that copies one database's information (db2) to anther database (sql server). The order of operations is very simple:
Check to see if anything has been updated in a certain time frame
Grab everything from the first database that is within the designated time frame
Map database information to POJOs
Divide subsets of POJOs into threads (pre defined # in properties file)
Threads cycle through each POJO Individually
Update the second database
I have everything working just fine, but at certain times of the day there is a huge jump in the amount of updates that need to take place (can get in to the hundreds of thousands).
Below you can see a generic version of my code. It follows the basic algorithm of the application. Object is generic, the actual application has 5 different types of specified objects each with its own updater thread class. But the generic functions below are exactly what they all look like. And in the updateDatabase() method, they all get added to threads and all run at the same time.
private void updateDatabase()
{
List<Thread> threads = new ArrayList<>();
addObjectThreads( threads );
startThreads( threads );
joinAllThreads( threads );
}
private void addObjectThreads( List<Thread> threads )
{
List<Object> objects = getTransformService().getObjects();
logger.info( "Found " + objects.size() + " Objects" );
createThreads( threads, objects, ObjectUpdaterThread.class );
}
private void createThreads( List<Thread> threads, List<?> objects, Class threadClass )
{
final int BASE_OBJECT_LOAD = 1;
int objectLoad = objects.size() / Database.getMaxThreads() > 0 ? objects.size() / Database.getMaxThreads() + BASE_OBJECT_LOAD : BASE_OBJECT_LOAD;
for (int i = 0; i < (objects.size() / objectLoad); ++i)
{
int startIndex = i * objectLoad;
int endIndex = (i + 1) * objectLoad;
try
{
List<?> objectSubList = objects.subList( startIndex, endIndex > objects.size() ? objects.size() : endIndex );
threads.add( new Thread( (Thread) threadClass.getConstructor( List.class ).newInstance( objectSubList ) ) );
}
catch (Exception exception)
{
logger.error( exception.getMessage() );
}
}
}
public class ObjectUpdaterThread extends BaseUpdaterThread
{
private List<Object> objects;
final private Logger logger = Logger.getLogger( ObjectUpdaterThread.class );
public ObjectUpdaterThread( List<Object> objects)
{
this.objects = objects;
}
public void run()
{
for (Object object : objects)
{
logger.info( "Now Updating Object: " + object.getId() );
getTransformService().updateObject( object );
}
}
}
All of these go to a spring service that looks like the code below. Again its generic, but each type of object has the exact same type of logic to them. The getObjects() from the code above are just one line pass throughs to the DAO so no need to really post that.
#Service
#Scope(value = "prototype")
public class TransformServiceImpl implements TransformService
{
final private Logger logger = Logger.getLogger( TransformServiceImpl.class );
#Autowired
private TransformDao transformDao;
#Override
public void updateObject( Object object )
{
String sql;
if ( object.exists() )
{
sql = Object.Mapper.UPDATE;
}
else
{
sql = Object.Mapper.INSERT;
}
boolean isCompleted = false;
while ( !isCompleted )
{
try
{
transformDao.updateObject( object, sql );
isCompleted = true;
}
catch (Exception exception)
{
logger.error( exception.getMessage() );
threadSleep();
logger.info( "Now retrying update for Object: " + object.getId() );
}
}
logger.info( "Updated Object: " + object.getId() );
}
}
Finally these all go to the DAO that looks like this:
#Repository
#Scope(value = "prototype")
public class TransformDaoImpl implements TransformDao
{
//#Resource is like #Autowired but with the added option of being able to specify the name
//Good for autowiring two different instances of the same class [NamedParameterJdbcTemplate]
//Another alternative = #Autowired #Qualifier(BEAN_NAME)
#Resource(name = "db2")
private NamedParameterJdbcTemplate db2;
#Resource(name = "sqlServer")
private NamedParameterJdbcTemplate sqlServer;
final private Logger logger = Logger.getLogger( TransformerImpl.class );
#Override
public void updateObject( Objet object, String sql )
{
MapSqlParameterSource source = new MapSqlParameterSource();
source.addValue( "column1_value", object.getColumn1Value() );
//put all source values from the POJO in just like above
sqlServer.update( sql, source );
}
}
My insert statements look like this:
"INSERT INTO dbo.OBJECT_TABLE " +
"(COLUMN1, COLUMN2...) " +
"VALUES(:column1_value, :column2_value... "
And my update statements look like this:
"UPDATE dbo.OBJECT_TABLE SET " +
"COLUMN1 = :column1_value, COLUMN2 = :column2_value, " +
"WHERE PRIMARY_KEY_COLUMN = :primary_key_value"
Its a lot of code and stuff I know, But I just wanted to layout everything I have in hopes that I can get help making this faster or more efficient. It takes hours on hours to update so many rows and it would nice if it only took a couple/few hours instead hours on hours. Thanks for any help. I welcome all learning experiences about spring, threads and databases.
If you're sending large amounts of SQL to the server, you should consider Batching it using the Statement.addBatch and Statement.executeBatch methods. The batches are finite in size (I always limited mine to 64K of SQL), but they dramatically lower the round trips to the database.
As I was iterating and creating SQL, I would keep track of how much I had batched already, when the SQL crossed the 64K boundary, I'd fire off an executeBatch and start a fresh one.
You may want to experiment with the 64K number, it may have been an Oracle limitation, which I was using at the time.
I can't speak to Spring, but batching is a part of the JDBC Statement. I'm sure it's straightforward to get to this.
Check to see if anything has been updated in a certain time frame
Grab everything from the first database that is within the designated time frame
Is there an index on the LAST_UPDATED_DATE column (or whatever you're using) in the source table? Rather than put the burden on your application, if it's within your control, why not write some triggers in the source database that create entries in an "update log" table? That way, all that your app would need to do is consume and execute those entries.
How are you managing your transactions? If you're creating a new transaction for each operation it's going to be brutally slow.
Regarding the threading code, have you considered using something more standard rather than writing your own? What you have is a pretty typical producer/consumer and Java has excellent support for that type of thing with ThreadPoolExecutor and numerous queue implementations to move data between threads that perform different tasks.
The benefit with using something off the shelf is that 1) it's well tested 2) there are numerous tuning options and sizing strategies that you can adjust to increase performance.
Also, rather than use 5 different thread types for each type of object that needs to be processed, have you considered encapsulating the processing logic for each type into separate strategy classes? That way, you could use a single pool of worker threads (which would be easier to size and tune).

How to implement a FSM - Finite State Machine in Java

I have something to do for work and I need your help.
We want to implement a FSM - Finite State Machine, to identify char sequence(like: A, B, C, A, C), and tell if it accepted.
We think to implement three classes: State, Event and Machine.
The state class presents a node in the FSM, we thought to implement it with State design pattern, every node will extend from the abstract class state and every class would handle different types of events and indicate transitions to a new state. Is it good idea in your opinion?
Second thing, we don't know how to save all the transitions. Again we thought to implement it with some kind of map, that hold the starting point and gets some kind of vector with the next states, but I'm not sure thats a good idea.
I would be happy to get some ideas of how to implement it or maybe you can give me some starting points.
How should I save the FSM, meaning how should I build the tree at the beginning of the program?
I googled it and found a lot of examples but nothing that helps me.
Thanks a lot.
The heart of a state machine is the transition table, which takes a state and a symbol (what you're calling an event) to a new state. That's just a two-index array of states. For sanity and type safety, declare the states and symbols as enumerations. I always add a "length" member in some way (language-specific) for checking array bounds. When I've hand-coded FSM's, I format the code in row and column format with whitespace fiddling. The other elements of a state machine are the initial state and the set of accepting states. The most direct implementation of the set of accepting states is an array of booleans indexed by the states. In Java, however, enumerations are classes, and you can specify an argument "accepting" in the declaration for each enumerated value and initialize it in the constructor for the enumeration.
For the machine type, you can write it as a generic class. It would take two type arguments, one for the states and one for the symbols, an array argument for the transition table, a single state for the initial. The only other detail (though it's critical) is that you have to call Enum.ordinal() to get an integer suitable for indexing the transition array, since you there's no syntax for directly declaring an array with a enumeration index (though there ought to be).
To preempt one issue, EnumMap won't work for the transition table, because the key required is a pair of enumeration values, not a single one.
enum State {
Initial( false ),
Final( true ),
Error( false );
static public final Integer length = 1 + Error.ordinal();
final boolean accepting;
State( boolean accepting ) {
this.accepting = accepting;
}
}
enum Symbol {
A, B, C;
static public final Integer length = 1 + C.ordinal();
}
State transition[][] = {
// A B C
{
State.Initial, State.Final, State.Error
}, {
State.Final, State.Initial, State.Error
}
};
EasyFSM is a dynamic Java Library which can be used to implement an FSM.
You can find documentation for the same at :
Finite State Machine in Java
Also, you can download the library at :
Java FSM Library : DynamicEasyFSM
You can implement Finite State Machine in two different ways.
Option 1:
Finite State machine with a pre-defined workflow : Recommended if you know all states in advance and state machine is almost fixed without any changes in future
Identify all possible states in your application
Identify all the events in your application
Identify all the conditions in your application, which may lead state transition
Occurrence of an event may cause transitions of state
Build a finite state machine by deciding a workflow of states & transitions.
e.g If an event 1 occurs at State 1, the state will be updated and machine state may still be in state 1.
If an event 2 occurs at State 1, on some condition evaluation, the system will move from State 1 to State 2
This design is based on State and Context patterns.
Have a look at Finite State Machine prototype classes.
Option 2:
Behavioural trees: Recommended if there are frequent changes to state machine workflow. You can dynamically add new behaviour without breaking the tree.
The base Task class provides a interface for all these tasks, the leaf tasks are the ones just mentioned, and the parent tasks are the interior nodes that decide which task to execute next.
The Tasks have only the logic they need to actually do what is required of them, all the decision logic of whether a task has started or not, if it needs to update, if it has finished with success, etc. is grouped in the TaskController class, and added by composition.
The decorators are tasks that “decorate” another class by wrapping over it and giving it additional logic.
Finally, the Blackboard class is a class owned by the parent AI that every task has a reference to. It works as a knowledge database for all the leaf tasks
Have a look at this article by Jaime Barrachina Verdia for more details
Hmm, I would suggest that you use Flyweight to implement the states. Purpose: Avoid the memory overhead of a large number of small objects. State machines can get very, very big.
http://en.wikipedia.org/wiki/Flyweight_pattern
I'm not sure that I see the need to use design pattern State to implement the nodes. The nodes in a state machine are stateless. They just match the current input symbol to the available transitions from the current state. That is, unless I have entirely forgotten how they work (which is a definite possiblilty).
If I were coding it, I would do something like this:
interface FsmNode {
public boolean canConsume(Symbol sym);
public FsmNode consume(Symbol sym);
// Other methods here to identify the state we are in
}
List<Symbol> input = getSymbols();
FsmNode current = getStartState();
for (final Symbol sym : input) {
if (!current.canConsume(sym)) {
throw new RuntimeException("FSM node " + current + " can't consume symbol " + sym);
}
current = current.consume(sym);
}
System.out.println("FSM consumed all input, end state is " + current);
What would Flyweight do in this case? Well, underneath the FsmNode there would probably be something like this:
Map<Integer, Map<Symbol, Integer>> fsm; // A state is an Integer, the transitions are from symbol to state number
FsmState makeState(int stateNum) {
return new FsmState() {
public FsmState consume(final Symbol sym) {
final Map<Symbol, Integer> transitions = fsm.get(stateNum);
if (transisions == null) {
throw new RuntimeException("Illegal state number " + stateNum);
}
final Integer nextState = transitions.get(sym); // May be null if no transition
return nextState;
}
public boolean canConsume(final Symbol sym) {
return consume(sym) != null;
}
}
}
This creates the State objects on a need-to-use basis, It allows you to use a much more efficient underlying mechanism to store the actual state machine. The one I use here (Map(Integer, Map(Symbol, Integer))) is not particulary efficient.
Note that the Wikipedia page focuses on the cases where many somewhat similar objects share the similar data, as is the case in the String implementation in Java. In my opinion, Flyweight is a tad more general, and covers any on-demand creation of objects with a short life span (use more CPU to save on a more efficient underlying data structure).
Consider the easy, lightweight Java library EasyFlow. From their docs:
With EasyFlow you can:
implement complex logic but keep your code simple and clean
handle asynchronous calls with ease and elegance
avoid concurrency by using event-driven programming approach
avoid StackOverflow error by avoiding recursion
simplify design, programming and testing of complex java applications
I design & implemented a simple finite state machine example with java.
IFiniteStateMachine: The public interface to manage the finite state machine
such as add new states to the finite state machine or transit to next states by specific actions.
interface IFiniteStateMachine {
void setStartState(IState startState);
void setEndState(IState endState);
void addState(IState startState, IState newState, Action action);
void removeState(String targetStateDesc);
IState getCurrentState();
IState getStartState();
IState getEndState();
void transit(Action action);
}
IState: The public interface to get state related info
such as state name and mappings to connected states.
interface IState {
// Returns the mapping for which one action will lead to another state
Map<String, IState> getAdjacentStates();
String getStateDesc();
void addTransit(Action action, IState nextState);
void removeTransit(String targetStateDesc);
}
Action: the class which will cause the transition of states.
public class Action {
private String mActionName;
public Action(String actionName) {
mActionName = actionName;
}
String getActionName() {
return mActionName;
}
#Override
public String toString() {
return mActionName;
}
}
StateImpl: the implementation of IState. I applied data structure such as HashMap to keep Action-State mappings.
public class StateImpl implements IState {
private HashMap<String, IState> mMapping = new HashMap<>();
private String mStateName;
public StateImpl(String stateName) {
mStateName = stateName;
}
#Override
public Map<String, IState> getAdjacentStates() {
return mMapping;
}
#Override
public String getStateDesc() {
return mStateName;
}
#Override
public void addTransit(Action action, IState state) {
mMapping.put(action.toString(), state);
}
#Override
public void removeTransit(String targetStateDesc) {
// get action which directs to target state
String targetAction = null;
for (Map.Entry<String, IState> entry : mMapping.entrySet()) {
IState state = entry.getValue();
if (state.getStateDesc().equals(targetStateDesc)) {
targetAction = entry.getKey();
}
}
mMapping.remove(targetAction);
}
}
FiniteStateMachineImpl: Implementation of IFiniteStateMachine. I use ArrayList to keep all the states.
public class FiniteStateMachineImpl implements IFiniteStateMachine {
private IState mStartState;
private IState mEndState;
private IState mCurrentState;
private ArrayList<IState> mAllStates = new ArrayList<>();
private HashMap<String, ArrayList<IState>> mMapForAllStates = new HashMap<>();
public FiniteStateMachineImpl(){}
#Override
public void setStartState(IState startState) {
mStartState = startState;
mCurrentState = startState;
mAllStates.add(startState);
// todo: might have some value
mMapForAllStates.put(startState.getStateDesc(), new ArrayList<IState>());
}
#Override
public void setEndState(IState endState) {
mEndState = endState;
mAllStates.add(endState);
mMapForAllStates.put(endState.getStateDesc(), new ArrayList<IState>());
}
#Override
public void addState(IState startState, IState newState, Action action) {
// validate startState, newState and action
// update mapping in finite state machine
mAllStates.add(newState);
final String startStateDesc = startState.getStateDesc();
final String newStateDesc = newState.getStateDesc();
mMapForAllStates.put(newStateDesc, new ArrayList<IState>());
ArrayList<IState> adjacentStateList = null;
if (mMapForAllStates.containsKey(startStateDesc)) {
adjacentStateList = mMapForAllStates.get(startStateDesc);
adjacentStateList.add(newState);
} else {
mAllStates.add(startState);
adjacentStateList = new ArrayList<>();
adjacentStateList.add(newState);
}
mMapForAllStates.put(startStateDesc, adjacentStateList);
// update mapping in startState
for (IState state : mAllStates) {
boolean isStartState = state.getStateDesc().equals(startState.getStateDesc());
if (isStartState) {
startState.addTransit(action, newState);
}
}
}
#Override
public void removeState(String targetStateDesc) {
// validate state
if (!mMapForAllStates.containsKey(targetStateDesc)) {
throw new RuntimeException("Don't have state: " + targetStateDesc);
} else {
// remove from mapping
mMapForAllStates.remove(targetStateDesc);
}
// update all state
IState targetState = null;
for (IState state : mAllStates) {
if (state.getStateDesc().equals(targetStateDesc)) {
targetState = state;
} else {
state.removeTransit(targetStateDesc);
}
}
mAllStates.remove(targetState);
}
#Override
public IState getCurrentState() {
return mCurrentState;
}
#Override
public void transit(Action action) {
if (mCurrentState == null) {
throw new RuntimeException("Please setup start state");
}
Map<String, IState> localMapping = mCurrentState.getAdjacentStates();
if (localMapping.containsKey(action.toString())) {
mCurrentState = localMapping.get(action.toString());
} else {
throw new RuntimeException("No action start from current state");
}
}
#Override
public IState getStartState() {
return mStartState;
}
#Override
public IState getEndState() {
return mEndState;
}
}
example:
public class example {
public static void main(String[] args) {
System.out.println("Finite state machine!!!");
IState startState = new StateImpl("start");
IState endState = new StateImpl("end");
IFiniteStateMachine fsm = new FiniteStateMachineImpl();
fsm.setStartState(startState);
fsm.setEndState(endState);
IState middle1 = new StateImpl("middle1");
middle1.addTransit(new Action("path1"), endState);
fsm.addState(startState, middle1, new Action("path1"));
System.out.println(fsm.getCurrentState().getStateDesc());
fsm.transit(new Action(("path1")));
System.out.println(fsm.getCurrentState().getStateDesc());
fsm.addState(middle1, endState, new Action("path1-end"));
fsm.transit(new Action(("path1-end")));
System.out.println(fsm.getCurrentState().getStateDesc());
fsm.addState(endState, middle1, new Action("path1-end"));
}
}
Full example on Github
Well this is an old question but while nobody mentioned here, I will advice to check two existing frameworks before you implement you own State Machines.
One is Spring State Machine most of you are familiar with Spring framework, which allow us to use several features of Spring like dependency injection and everything else that Spring can offer.
It is really great for modelling the lifecycle of an Apparat, with states like INITIALIZING, STARTED, ERROR, RECOVERING, SHUTTINGDOWN, etc.. but I see lots of people are trying to model a Shopping Chart, a Reservation System with it, the memory footprint a Spring State Machine is relatively big to model millions of Shopping Charts or Reservations.
One other disadvantage, Spring State Machine, while has a capability to persist itself for long running processes but it does not have any mechanism to adapt to changes in these processes, if you persist a process and you have to recover it lets say 10 days later with a change occurred in your business process because of new software release / requirement, you have no built in means to deal with it.
I have several blogs, blog1 blog2, demonstrating how you can program Spring State Machine, specially model driven way, if you want to check it.
Mainly because the disadvantages I mentioned, I advice you to look another framework first, Akka FSM (Finite State Machine) which is more fitting with its low memory footprint to have millions and millions of instances and has a capability to adapt changing long running processes.
Now you can develop with Akka framework with Java but believe me because of some missing language elements, you don't want to read the produced code, Scala is a much more fitting language to develop with Akka. Now I hear you saying Scala is too complex, I can't convince my project leads to develop with Scala, to convince you all this is an option, I developed a Proof of Concept application using a Java/Scala hybrid with all Scala Akka Finite State Machine code generated from an UML model, if you want to check it out here the links to the blogs, blog3 blog4.
I hope this information would help you.
Here is a SUPER SIMPLE implementation/example of a FSM using just "if-else"s which avoids all of the above subclassing answers (taken from Using Finite State Machines for Pattern Matching in Java, where he is looking for a string which ends with "#" followed by numbers followed by "#"--see state graph here):
public static void main(String[] args) {
String s = "A1#312#";
String digits = "0123456789";
int state = 0;
for (int ind = 0; ind < s.length(); ind++) {
if (state == 0) {
if (s.charAt(ind) == '#')
state = 1;
} else {
boolean isNumber = digits.indexOf(s.charAt(ind)) != -1;
if (state == 1) {
if (isNumber)
state = 2;
else if (s.charAt(ind) == '#')
state = 1;
else
state = 0;
} else if (state == 2) {
if (s.charAt(ind) == '#') {
state = 3;
} else if (isNumber) {
state = 2;
} else if (s.charAt(ind) == '#')
state = 1;
else
state = 0;
} else if (state == 3) {
if (s.charAt(ind) == '#')
state = 1;
else
state = 0;
}
}
} //end for loop
if (state == 3)
System.out.println("It matches");
else
System.out.println("It does not match");
}
P.S: Does not answer your question directly, but shows you how to implement a FSM very easily in Java.

Categories

Resources