Is it possible to use generic types with declare parents such that a class defined with generics implements an interface with the same generic types
i.e declare parents: AClass<Generic1,Generic2> implements
AnInterface<Generic1,Generic2>
What I am saying is whether it is possible to pass the generic types of the child to parents
Kind of. Check this out:
Generic class:
package de.scrum_master.app;
public class KeyValuePair<K,V> {
private K key;
private V value;
KeyValuePair(K key, V value) {
this.key = key;
this.value = value;
}
public K getKey() { return key; }
public V getValue() { return value; }
#Override
public String toString() {
return "KeyValuePair [key=" + key + ", value=" + value + "]";
}
}
Generic interface:
package de.scrum_master.app;
public interface KeyValueComparator<K, V> {
boolean equalsKey(K otherKey);
boolean equalsValue(V otherValue);
}
Aspect with ITD:
The ITD (inter-type definition) makes sure that the class implements the interface and also gets method implementations for the interface at the same time.
package de.scrum_master.aspect;
import de.scrum_master.app.KeyValuePair;
import de.scrum_master.app.KeyValueComparator;
public aspect InterfaceIntroductionAspect {
declare parents : KeyValuePair implements KeyValueComparator;
public boolean KeyValuePair.equalsKey(K otherKey) {
return this.getKey().equals(otherKey);
}
public boolean KeyValuePair.equalsValue(V otherValue) {
return this.getValue().equals(otherValue);
}
}
Driver application:
Create two different types of class objects and try to cast them both to the introduced interface:
package de.scrum_master.app;
public class Application {
public static void main(String[] args) {
KeyValuePair<Integer, String> pair1 = new KeyValuePair<>(11, "eleven");
System.out.println(pair1);
KeyValueComparator<Integer, String> comparator1 = (KeyValueComparator<Integer, String>) pair1;
System.out.println("equalsKey = " + comparator1.equalsKey(12));
System.out.println("equalsValue = " + comparator1.equalsValue("eleven"));
KeyValuePair<String, Integer> pair2 = new KeyValuePair<>("twelve", 12);
System.out.println(pair2);
KeyValueComparator<String, Integer> comparator2 = (KeyValueComparator<String, Integer>) pair2;
System.out.println("equalsKey = " + comparator2.equalsKey("twelve"));
System.out.println("equalsValue = " + comparator2.equalsValue(11));
}
}
Output:
KeyValuePair [key=11, value=eleven]
equalsKey = false
equalsValue = true
KeyValuePair [key=twelve, value=12]
equalsKey = true
equalsValue = false
Related
I've got some configuration values in a JSON file which I want to parse via gson to a data-class. I want to generate a new class, based on the created data-class where the values are final.
This all should happen during my CI-Pipeline and the generated class should then be used when my application is running.
Simple example to clarify:
I've got this data class
data class MyDataClass(val name:String, val age:Int)
and via parsing (gson) a instance like this is created
MyDataClass("john", 42)
Is there a way to create a (data) class based on the new instance of MyDataClass so anything like this will be created?
class MyDataClassFinal{
val name = "john"
val age = 42
}
Use .copy() and modify only the parameters you need to. For example:
val joe = MyDataClass(“Joe”, 42)
val mary = joe.copy(name = “Mary”) // age is 42
I've got some configuration values in a json file wich I want to parse via gson to a data-class. I want to generate a new class, based on the created data-class where the values are final.
There is nothing for you to do here. The data class, as you've described it, is final. It is not open so the class is final and the fields are vals and only set via constructor, so they can't be changed, so they too are final.
You can see the Java equivalent of the class by going to doing a search for "Actions", looking for Kotlin Bytecode, then hit Decompile to see the Java source. It looks like this:
#Metadata(
mv = {1, 1, 18},
bv = {1, 0, 3},
k = 1,
d1 = {"\u0000 \n\u0002\u0018\u0002\n\u0002\u0010\u0000\n\u0000\n\u0002\u0010\u000e\n\u0000\n\u0002\u0010\b\n\u0002\b\t\n\u0002\u0010\u000b\n\u0002\b\u0004\b\u0086\b\u0018\u00002\u00020\u0001B\u0015\u0012\u0006\u0010\u0002\u001a\u00020\u0003\u0012\u0006\u0010\u0004\u001a\u00020\u0005¢\u0006\u0002\u0010\u0006J\t\u0010\u000b\u001a\u00020\u0003HÆ\u0003J\t\u0010\f\u001a\u00020\u0005HÆ\u0003J\u001d\u0010\r\u001a\u00020\u00002\b\b\u0002\u0010\u0002\u001a\u00020\u00032\b\b\u0002\u0010\u0004\u001a\u00020\u0005HÆ\u0001J\u0013\u0010\u000e\u001a\u00020\u000f2\b\u0010\u0010\u001a\u0004\u0018\u00010\u0001HÖ\u0003J\t\u0010\u0011\u001a\u00020\u0005HÖ\u0001J\t\u0010\u0012\u001a\u00020\u0003HÖ\u0001R\u0011\u0010\u0004\u001a\u00020\u0005¢\u0006\b\n\u0000\u001a\u0004\b\u0007\u0010\bR\u0011\u0010\u0002\u001a\u00020\u0003¢\u0006\b\n\u0000\u001a\u0004\b\t\u0010\n¨\u0006\u0013"},
d2 = {"Lcore/lib/extensions/MyDataClass;", "", "name", "", "age", "", "(Ljava/lang/String;I)V", "getAge", "()I", "getName", "()Ljava/lang/String;", "component1", "component2", "copy", "equals", "", "other", "hashCode", "toString", "treking-android.dominicore-android"}
)
public final class MyDataClass {
#NotNull
private final String name;
private final int age;
#NotNull
public final String getName() {
return this.name;
}
public final int getAge() {
return this.age;
}
public MyDataClass(#NotNull String name, int age) {
Intrinsics.checkParameterIsNotNull(name, "name");
super();
this.name = name;
this.age = age;
}
#NotNull
public final String component1() {
return this.name;
}
public final int component2() {
return this.age;
}
#NotNull
public final MyDataClass copy(#NotNull String name, int age) {
Intrinsics.checkParameterIsNotNull(name, "name");
return new MyDataClass(name, age);
}
// $FF: synthetic method
public static MyDataClass copy$default(MyDataClass var0, String var1, int var2, int var3, Object var4) {
if ((var3 & 1) != 0) {
var1 = var0.name;
}
if ((var3 & 2) != 0) {
var2 = var0.age;
}
return var0.copy(var1, var2);
}
#NotNull
public String toString() {
return "MyDataClass(name=" + this.name + ", age=" + this.age + ")";
}
public int hashCode() {
String var10000 = this.name;
return (var10000 != null ? var10000.hashCode() : 0) * 31 + this.age;
}
public boolean equals(#Nullable Object var1) {
if (this != var1) {
if (var1 instanceof MyDataClass) {
MyDataClass var2 = (MyDataClass)var1;
if (Intrinsics.areEqual(this.name, var2.name) && this.age == var2.age) {
return true;
}
}
return false;
} else {
return true;
}
}
}
As you can see, the class, it's fields, and their accessors are all final.
I'm trying to serialize an object (Root), with some duplicated entries of MyObject. Just want store the whole objects one, I'm using #JsonIdentityReference, which works pretty well.
However, I realize that it will generate un-deserializable object, if there're equal objects with different reference. I wonder if there's a configuration in Jackson to change this behavior, thanks!
#Value
#AllArgsConstructor
#NoArgsConstructor(force = true)
class Root {
private List<MyObject> allObjects;
private Map<String, MyObject> objectMap;
}
#Value
#AllArgsConstructor
#NoArgsConstructor(force = true)
#JsonIdentityReference
#JsonIdentityInfo(generator = ObjectIdGenerators.PropertyGenerator.class, property = "id")
class MyObject {
private String id;
private int value;
}
public class Main {
public static void main() throws JsonProcessingException {
// Constructing equal objects
val obj1 = new MyObject("a", 1);
val obj2 = new MyObject("a", 1);
assert obj1.equals(obj2);
val root = new Root(
Lists.newArrayList(obj1),
ImmutableMap.of(
"lorem", obj2
)
);
val objectMapper = new ObjectMapper();
val json = objectMapper.writeValueAsString(root);
// {"allObjects":[{"id":"a","value":1}],"objectMap":{"lorem":{"id":"a","value":1}}}
// Note here both obj1 and obj2 are expanded.
// Exception: Already had POJO for id
val deserialized = objectMapper.readValue(json, Root.class);
assert root.equals(deserialized);
}
}
I'm using Jackson 2.10.
Full stacktrace:
Exception in thread "main" com.fasterxml.jackson.databind.JsonMappingException: Already had POJO for id (java.lang.String) [[ObjectId: key=a, type=com.fasterxml.jackson.databind.deser.impl.PropertyBasedObjectIdGenerator, scope=java.lang.Object]] (through reference chain: Root["objectMap"]->java.util.LinkedHashMap["lorem"]->MyObject["id"])
at com.fasterxml.jackson.databind.JsonMappingException.wrapWithPath(JsonMappingException.java:394)
at com.fasterxml.jackson.databind.JsonMappingException.wrapWithPath(JsonMappingException.java:353)
at com.fasterxml.jackson.databind.deser.BeanDeserializerBase.wrapAndThrow(BeanDeserializerBase.java:1714)
at com.fasterxml.jackson.databind.deser.BeanDeserializer.deserializeFromObject(BeanDeserializer.java:371)
at com.fasterxml.jackson.databind.deser.BeanDeserializerBase.deserializeWithObjectId(BeanDeserializerBase.java:1257)
at com.fasterxml.jackson.databind.deser.BeanDeserializer.deserialize(BeanDeserializer.java:157)
at com.fasterxml.jackson.databind.deser.std.MapDeserializer._readAndBindStringKeyMap(MapDeserializer.java:527)
at com.fasterxml.jackson.databind.deser.std.MapDeserializer.deserialize(MapDeserializer.java:364)
at com.fasterxml.jackson.databind.deser.std.MapDeserializer.deserialize(MapDeserializer.java:29)
at com.fasterxml.jackson.databind.deser.impl.FieldProperty.deserializeAndSet(FieldProperty.java:138)
at com.fasterxml.jackson.databind.deser.BeanDeserializer.vanillaDeserialize(BeanDeserializer.java:288)
at com.fasterxml.jackson.databind.deser.BeanDeserializer.deserialize(BeanDeserializer.java:151)
at com.fasterxml.jackson.databind.ObjectMapper._readMapAndClose(ObjectMapper.java:4202)
at com.fasterxml.jackson.databind.ObjectMapper.readValue(ObjectMapper.java:3205)
at com.fasterxml.jackson.databind.ObjectMapper.readValue(ObjectMapper.java:3173)
at Main.main(Main.java:53)
Caused by: java.lang.IllegalStateException: Already had POJO for id (java.lang.String) [[ObjectId: key=a, type=com.fasterxml.jackson.databind.deser.impl.PropertyBasedObjectIdGenerator, scope=java.lang.Object]]
at com.fasterxml.jackson.annotation.SimpleObjectIdResolver.bindItem(SimpleObjectIdResolver.java:24)
at com.fasterxml.jackson.databind.deser.impl.ReadableObjectId.bindItem(ReadableObjectId.java:57)
at com.fasterxml.jackson.databind.deser.impl.ObjectIdValueProperty.deserializeSetAndReturn(ObjectIdValueProperty.java:101)
at com.fasterxml.jackson.databind.deser.impl.ObjectIdValueProperty.deserializeAndSet(ObjectIdValueProperty.java:83)
at com.fasterxml.jackson.databind.deser.BeanDeserializer.deserializeFromObject(BeanDeserializer.java:369)
... 14 more
As I mentioned earlier, this setup only works if obj1 == obj2, as the two objects with same ID should be identity-equal. In that case, the second object would also net get expanded during serialization (alwaysAsId = false only expands the first object).
However, if you want to have this setup and are fine with the serialization, you could use a custom Resolver for deserialization that stores a single instance per key:
#JsonIdentityReference(alwaysAsId = false)
#JsonIdentityInfo(generator = ObjectIdGenerators.PropertyGenerator.class, property = "id", resolver = CustomScopeResolver.class)
static class MyObject {
private String id;
// ...
}
class CustomScopeResolver implements ObjectIdResolver {
Map<String, MyObject> data = new HashMap<>();
#Override
public void bindItem(final IdKey id, final Object pojo) {
data.put(id.key.toString(), (MyObject) pojo);
}
#Override
public Object resolveId(final IdKey id) {
return data.get(id.key);
}
#Override
public ObjectIdResolver newForDeserialization(final Object context) {
return new CustomScopeResolver();
}
#Override
public boolean canUseFor(final ObjectIdResolver resolverType) {
return false;
}
}
NEW EDIT: Apparently, its very easy: Just turn on objectMapper.configure(SerializationFeature.USE_EQUALITY_FOR_OBJECT_ID, true); so that the DefaultSerializerProvider uses a regular Hashmap instead of an IdentityHashMap to manage the serialized beans.
DEPRECATED: Update for Serialization: It is possible to achieve this by adding a custom SerializationProvider:
class CustomEqualObjectsSerializerProvider extends DefaultSerializerProvider {
private final Collection<MyObject> data = new HashSet<>();
private final SerializerProvider src;
private final SerializationConfig config;
private final SerializerFactory f;
public CustomEqualObjectsSerializerProvider(
final SerializerProvider src,
final SerializationConfig config,
final SerializerFactory f) {
super(src, config, f);
this.src = src;
this.config = config;
this.f = f;
}
#Override
public DefaultSerializerProvider createInstance(final SerializationConfig config, final SerializerFactory jsf) {
return new CustomEqualObjectsSerializerProvider(src, this.config, f);
}
#Override
public WritableObjectId findObjectId(final Object forPojo, final ObjectIdGenerator<?> generatorType) {
// check if there is an equivalent pojo, use it if exists
final Optional<MyObject> equivalentObject = data.stream()
.filter(forPojo::equals)
.findFirst();
if (equivalentObject.isPresent()) {
return super.findObjectId(equivalentObject.get(), generatorType);
} else {
if (forPojo instanceof MyObject) {
data.add((MyObject) forPojo);
}
return super.findObjectId(forPojo, generatorType);
}
}
}
#Test
public void main() throws IOException {
// Constructing equal objects
final MyObject obj1 = new MyObject();
obj1.setId("a");
final MyObject obj2 = new MyObject();
obj2.setId("a");
assert obj1.equals(obj2);
final Root root = new Root();
root.setAllObjects(Collections.singletonList(obj1));
root.setObjectMap(Collections.singletonMap(
"lorem", obj2));
final ObjectMapper objectMapper = new ObjectMapper();
objectMapper.setSerializerProvider(
new CustomEqualObjectsSerializerProvider(
objectMapper.getSerializerProvider(),
objectMapper.getSerializationConfig(),
objectMapper.getSerializerFactory()));
final String json = objectMapper.writeValueAsString(root);
System.out.println(json); // second object is not expanded!
}
i've written a utility to monitor individual business transactions. For example, Alice calls a method which calls more methods and i want info on just Alice's call, separate from Bob's call to the same method.
Right now the entry point creates a Transaction object and it's passed as an argument to each method:
class Example {
public Item getOrderEntryPoint(int orderId) {
Transaction transaction = transactionManager.create();
transaction.trace("getOrderEntryPoint");
Order order = getOrder(orderId, transaction);
transaction.stop();
logger.info(transaction);
return item;
}
private Order getOrder(int orderId, Transaction t) {
t.trace("getOrder");
Order order = getItems(itemId, t);
t.addStat("number of items", order.getItems().size());
for (Item item : order.getItems()) {
SpecialOffer offer = getSpecialOffer(item, t);
if (null != offer) {
t.incrementStat("offers", 1);
}
}
t.stop();
return order;
}
private SpecialOffer getSpecialOffer(Item item, Transaction t) {
t.trace("getSpecialOffer(" + item.id + ")", TraceCategory.Database);
return offerRepository.getByItem(item);
t.stop();
}
}
This will print to the log something like:
Transaction started by Alice at 10:42
Statistics:
number of items : 3
offers : 1
Category Timings (longest first):
DB : 2s 903ms
code : 187ms
Timings (longest first):
getSpecialOffer(1013) : 626ms
getItems : 594ms
Trace:
getOrderEntryPoint (7ms)
getOrder (594ms)
getSpecialOffer(911) (90ms)
getSpecialOffer(1013) (626ms)
getSpecialOffer(2942) (113ms)
It works great but passing the transaction object around is ugly. Someone suggested AOP but i don't see how to pass the transaction created in the first method to all the other methods.
The Transaction object is pretty simple:
public class Transaction {
private String uuid = UUID.createRandom();
private List<TraceEvent> events = new ArrayList<>();
private Map<String,Int> stats = new HashMap<>();
}
class TraceEvent {
private String name;
private long durationInMs;
}
The app that uses it is a Web app, and this multi-threaded, but the individual transactions are on a single thread - no multi-threading, async code, competition for resources, etc.
My attempt at an annotation:
#Around("execution(* *(..)) && #annotation(Trace)")
public Object around(ProceedingJoinPoint point) {
String methodName = MethodSignature.class.cast(point.getSignature()).getMethod().getName();
//--- Where do i get this call's instance of TRANSACTION from?
if (null == transaction) {
transaction = TransactionManager.createTransaction();
}
transaction.trace(methodName);
Object result = point.proceed();
transaction.stop();
return result;
Introduction
Unfortunately, your pseudo code does not compile. It contains several syntactical and logical errors. Furthermore, some helper classes are missing. If I did not have spare time today and was looking for a puzzle to solve, I would not have bothered making my own MCVE out of it, because that would actually have been your job. Please do read the MCVE article and learn to create one next time, otherwise you will not get a lot of qualified help here. This was your free shot because you are new on SO.
Original situation: passing through transaction objects in method calls
Application helper classes:
package de.scrum_master.app;
public class Item {
private int id;
public Item(int id) {
this.id = id;
}
public int getId() {
return id;
}
#Override
public String toString() {
return "Item[id=" + id + "]";
}
}
package de.scrum_master.app;
public class SpecialOffer {}
package de.scrum_master.app;
public class OfferRepository {
public SpecialOffer getByItem(Item item) {
if (item.getId() < 30)
return new SpecialOffer();
return null;
}
}
package de.scrum_master.app;
import java.util.ArrayList;
import java.util.List;
public class Order {
private int id;
public Order(int id) {
this.id = id;
}
public List<Item> getItems() {
List<Item> items = new ArrayList<>();
int offset = id == 12345 ? 0 : 1;
items.add(new Item(11 + offset, this));
items.add(new Item(22 + offset, this));
items.add(new Item(33 + offset, this));
return items;
}
}
Trace classes:
package de.scrum_master.trace;
public enum TraceCategory {
Code, Database
}
package de.scrum_master.trace;
class TraceEvent {
private String name;
private TraceCategory category;
private long durationInMs;
private boolean finished = false;
public TraceEvent(String name, TraceCategory category, long startTime) {
this.name = name;
this.category = category;
this.durationInMs = startTime;
}
public long getDurationInMs() {
return durationInMs;
}
public void setDurationInMs(long durationInMs) {
this.durationInMs = durationInMs;
}
public boolean isFinished() {
return finished;
}
public void setFinished(boolean finished) {
this.finished = finished;
}
#Override
public String toString() {
return "TraceEvent[name=" + name + ", category=" + category +
", durationInMs=" + durationInMs + ", finished=" + finished + "]";
}
}
Transaction classes:
Here I tried to mimic your own Transaction class with as few as possible changes, but there was a lot I had to add and modify in order to emulate a simplified version of your trace output. This is not thread-safe and the way I am locating the last unfinished TraceEvent is not nice and only works cleanly if there are not exceptions. But you get the idea, I hope. The point is to just make it basically work and subsequently get log output similar to your example. If this was originally my code, I would have solved it differently.
package de.scrum_master.trace;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
import java.util.UUID;
public class Transaction {
private String uuid = UUID.randomUUID().toString();
private List<TraceEvent> events = new ArrayList<>();
private Map<String, Integer> stats = new HashMap<>();
public void trace(String message) {
trace(message, TraceCategory.Code);
}
public void trace(String message, TraceCategory category) {
events.add(new TraceEvent(message, category, System.currentTimeMillis()));
}
public void stop() {
TraceEvent event = getLastUnfinishedEvent();
event.setDurationInMs(System.currentTimeMillis() - event.getDurationInMs());
event.setFinished(true);
}
private TraceEvent getLastUnfinishedEvent() {
return events
.stream()
.filter(event -> !event.isFinished())
.reduce((first, second) -> second)
.orElse(null);
}
public void addStat(String text, int size) {
stats.put(text, size);
}
public void incrementStat(String text, int increment) {
Integer currentCount = stats.get(text);
if (currentCount == null)
currentCount = 0;
stats.put(text, currentCount + increment);
}
#Override
public String toString() {
return "Transaction {" +
toStringUUID() +
toStringStats() +
toStringEvents() +
"\n}\n";
}
private String toStringUUID() {
return "\n uuid = " + uuid;
}
private String toStringStats() {
String result = "\n stats = {";
for (Entry<String, Integer> statEntry : stats.entrySet())
result += "\n " + statEntry;
return result + "\n }";
}
private String toStringEvents() {
String result = "\n events = {";
for (TraceEvent event : events)
result += "\n " + event;
return result + "\n }";
}
}
package de.scrum_master.trace;
public class TransactionManager {
public Transaction create() {
return new Transaction();
}
}
Example driver application:
package de.scrum_master.app;
import de.scrum_master.trace.TraceCategory;
import de.scrum_master.trace.Transaction;
import de.scrum_master.trace.TransactionManager;
public class Example {
private TransactionManager transactionManager = new TransactionManager();
private OfferRepository offerRepository = new OfferRepository();
public Order getOrderEntryPoint(int orderId) {
Transaction transaction = transactionManager.create();
transaction.trace("getOrderEntryPoint");
sleep(100);
Order order = getOrder(orderId, transaction);
transaction.stop();
System.out.println(transaction);
return order;
}
private Order getOrder(int orderId, Transaction t) {
t.trace("getOrder");
sleep(200);
Order order = new Order(orderId);
t.addStat("number of items", order.getItems().size());
for (Item item : order.getItems()) {
SpecialOffer offer = getSpecialOffer(item, t);
if (null != offer)
t.incrementStat("special offers", 1);
}
t.stop();
return order;
}
private SpecialOffer getSpecialOffer(Item item, Transaction t) {
t.trace("getSpecialOffer(" + item.getId() + ")", TraceCategory.Database);
sleep(50);
SpecialOffer specialOffer = offerRepository.getByItem(item);
t.stop();
return specialOffer;
}
private void sleep(long millis) {
try {
Thread.sleep(millis);
} catch (InterruptedException e) {
e.printStackTrace();
}
}
public static void main(String[] args) {
new Example().getOrderEntryPoint(12345);
new Example().getOrderEntryPoint(23456);
}
}
If you run this code, the output is as follows:
Transaction {
uuid = 62ec9739-bd32-4a56-b6b3-a8a13624961a
stats = {
special offers=2
number of items=3
}
events = {
TraceEvent[name=getOrderEntryPoint, category=Code, durationInMs=561, finished=true]
TraceEvent[name=getOrder, category=Code, durationInMs=451, finished=true]
TraceEvent[name=getSpecialOffer(11), category=Database, durationInMs=117, finished=true]
TraceEvent[name=getSpecialOffer(22), category=Database, durationInMs=69, finished=true]
TraceEvent[name=getSpecialOffer(33), category=Database, durationInMs=63, finished=true]
}
}
Transaction {
uuid = a420cd70-96e5-44c4-a0a4-87e421d05e87
stats = {
special offers=2
number of items=3
}
events = {
TraceEvent[name=getOrderEntryPoint, category=Code, durationInMs=469, finished=true]
TraceEvent[name=getOrder, category=Code, durationInMs=369, finished=true]
TraceEvent[name=getSpecialOffer(12), category=Database, durationInMs=53, finished=true]
TraceEvent[name=getSpecialOffer(23), category=Database, durationInMs=63, finished=true]
TraceEvent[name=getSpecialOffer(34), category=Database, durationInMs=53, finished=true]
}
}
AOP refactoring
Preface
Please note that I am using AspectJ here because two things about your code would never work with Spring AOP because it works with a delegation pattern based on dynamic proxies:
self-invocation (internally calling a method of the same class or super-class)
intercepting private methods
Because of these Spring AOP limitations I advise you to either refactor your code so as to avoid the two issues above or to configure your Spring applications to use full AspectJ via LTW (load-time weaving) instead.
As you noticed, my sample code does not use Spring at all because AspectJ is completely independent of Spring and works with any Java application (or other JVM languages, too).
Refactoring idea
Now what should you do in order to get rid of passing around tracing information (Transaction objects), polluting your core application code and tangling it with trace calls?
You extract transaction tracing into an aspect taking care of all trace(..) and stop() calls.
Unfortunately your Transaction class contains different types of information and does different things, so you cannot completely get rid of context information about how to trace for each affected method. But at least you can extract that context information from the method bodies and transform it into a declarative form using annotations with parameters.
These annotations can be targeted by an aspect taking care of handling transaction tracing.
Added and updated code, iteration 1
Annotations related to transaction tracing:
package de.scrum_master.trace;
import static java.lang.annotation.ElementType.METHOD;
import static java.lang.annotation.RetentionPolicy.RUNTIME;
import java.lang.annotation.Retention;
import java.lang.annotation.Target;
#Retention(RUNTIME)
#Target(METHOD)
public #interface TransactionEntryPoint {}
package de.scrum_master.trace;
import static java.lang.annotation.ElementType.METHOD;
import static java.lang.annotation.RetentionPolicy.RUNTIME;
import java.lang.annotation.Retention;
import java.lang.annotation.Target;
#Retention(RUNTIME)
#Target(METHOD)
public #interface TransactionTrace {
String message() default "__METHOD_NAME__";
TraceCategory category() default TraceCategory.Code;
String addStat() default "";
String incrementStat() default "";
}
Refactored application classes with annotations:
package de.scrum_master.app;
import java.util.ArrayList;
import java.util.List;
import de.scrum_master.trace.TransactionTrace;
public class Order {
private int id;
public Order(int id) {
this.id = id;
}
#TransactionTrace(message = "", addStat = "number of items")
public List<Item> getItems() {
List<Item> items = new ArrayList<>();
int offset = id == 12345 ? 0 : 1;
items.add(new Item(11 + offset));
items.add(new Item(22 + offset));
items.add(new Item(33 + offset));
return items;
}
}
Nothing much here, only added an annotation to getItems(). But the sample application class changes massively, getting much cleaner and simpler:
package de.scrum_master.app;
import de.scrum_master.trace.TraceCategory;
import de.scrum_master.trace.TransactionEntryPoint;
import de.scrum_master.trace.TransactionTrace;
public class Example {
private OfferRepository offerRepository = new OfferRepository();
#TransactionEntryPoint
#TransactionTrace
public Order getOrderEntryPoint(int orderId) {
sleep(100);
Order order = getOrder(orderId);
return order;
}
#TransactionTrace
private Order getOrder(int orderId) {
sleep(200);
Order order = new Order(orderId);
for (Item item : order.getItems()) {
SpecialOffer offer = getSpecialOffer(item);
// Do something with special offers
}
return order;
}
#TransactionTrace(category = TraceCategory.Database, incrementStat = "specialOffers")
private SpecialOffer getSpecialOffer(Item item) {
sleep(50);
SpecialOffer specialOffer = offerRepository.getByItem(item);
return specialOffer;
}
private void sleep(long millis) {
try {
Thread.sleep(millis);
} catch (InterruptedException e) {
e.printStackTrace();
}
}
public static void main(String[] args) {
new Example().getOrderEntryPoint(12345);
new Example().getOrderEntryPoint(23456);
}
}
See? Except for a few annotations there is nothing left of the transaction tracing logic, the application code only takes care of its core concern. If you also remove the sleep() method which only makes the application slower for demonstration purposes (because we want some nice statistics with measured times >0 ms), the class gets even more compact.
But of course we need to put the transaction tracing logic somewhere, more precisely modularise it into an AspectJ aspect:
Transaction tracing aspect:
package de.scrum_master.trace;
import java.lang.reflect.Array;
import java.util.Arrays;
import java.util.Collection;
import java.util.stream.Collectors;
import org.aspectj.lang.JoinPoint;
import org.aspectj.lang.ProceedingJoinPoint;
import org.aspectj.lang.annotation.After;
import org.aspectj.lang.annotation.Around;
import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.Pointcut;
import org.aspectj.lang.reflect.MethodSignature;
#Aspect("percflow(entryPoint())")
public class TransactionTraceAspect {
private static TransactionManager transactionManager = new TransactionManager();
private Transaction transaction = transactionManager.create();
#Pointcut("execution(* *(..)) && #annotation(de.scrum_master.trace.TransactionEntryPoint)")
private static void entryPoint() {}
#Around("execution(* *(..)) && #annotation(transactionTrace)")
public Object doTrace(ProceedingJoinPoint joinPoint, TransactionTrace transactionTrace) throws Throwable {
preTrace(transactionTrace, joinPoint);
Object result = joinPoint.proceed();
postTrace(transactionTrace);
addStat(transactionTrace, result);
incrementStat(transactionTrace, result);
return result;
}
private void preTrace(TransactionTrace transactionTrace, ProceedingJoinPoint joinPoint) {
String traceMessage = transactionTrace.message();
if ("".equals(traceMessage))
return;
MethodSignature signature = (MethodSignature) joinPoint.getSignature();
if ("__METHOD_NAME__".equals(traceMessage)) {
traceMessage = signature.getName() + "(";
traceMessage += Arrays.stream(joinPoint.getArgs()).map(arg -> arg.toString()).collect(Collectors.joining(", "));
traceMessage += ")";
}
transaction.trace(traceMessage, transactionTrace.category());
}
private void postTrace(TransactionTrace transactionTrace) {
if ("".equals(transactionTrace.message()))
return;
transaction.stop();
}
private void addStat(TransactionTrace transactionTrace, Object result) {
if ("".equals(transactionTrace.addStat()) || result == null)
return;
if (result instanceof Collection)
transaction.addStat(transactionTrace.addStat(), ((Collection<?>) result).size());
else if (result.getClass().isArray())
transaction.addStat(transactionTrace.addStat(), Array.getLength(result));
}
private void incrementStat(TransactionTrace transactionTrace, Object result) {
if ("".equals(transactionTrace.incrementStat()) || result == null)
return;
transaction.incrementStat(transactionTrace.incrementStat(), 1);
}
#After("entryPoint()")
public void logFinishedTransaction(JoinPoint joinPoint) {
System.out.println(transaction);
}
}
Let me explain what this aspect does:
#Pointcut(..) entryPoint() says: Find me all methods in the code annotated by #TransactionEntryPoint. This pointcut is used in two places:
#Aspect("percflow(entryPoint())") says: Create one aspect instance for each control flow beginning at a transaction entry point.
#After("entryPoint()") logFinishedTransaction(..) says: Execute this advice (AOP terminology for a method linked to a pointcut) after an entry point methods is finished. The corresponding method just prints the transaction statistics just like in the original code at the end of Example.getOrderEntryPoint(..).
#Around("execution(* *(..)) && #annotation(transactionTrace)") doTrace(..)says: Wrap methods annotated by TransactionTrace and do the following (method body):
add new trace element and start measuring time
execute original (wrapped) method and store result
update trace element with measured time
add one type of statistics (optional)
increment another type of statistics (optional)
return wrapped method's result to its caller
The private methods are just helpers for the #Around advice.
The console log when running the updated Example class and active AspectJ is:
Transaction {
uuid = 4529d325-c604-441d-8997-45ca659abb14
stats = {
specialOffers=2
number of items=3
}
events = {
TraceEvent[name=getOrderEntryPoint(12345), category=Code, durationInMs=468, finished=true]
TraceEvent[name=getOrder(12345), category=Code, durationInMs=366, finished=true]
TraceEvent[name=getSpecialOffer(Item[id=11]), category=Database, durationInMs=59, finished=true]
TraceEvent[name=getSpecialOffer(Item[id=22]), category=Database, durationInMs=50, finished=true]
TraceEvent[name=getSpecialOffer(Item[id=33]), category=Database, durationInMs=51, finished=true]
}
}
Transaction {
uuid = ef76a996-8621-478b-a376-e9f7a729a501
stats = {
specialOffers=2
number of items=3
}
events = {
TraceEvent[name=getOrderEntryPoint(23456), category=Code, durationInMs=452, finished=true]
TraceEvent[name=getOrder(23456), category=Code, durationInMs=351, finished=true]
TraceEvent[name=getSpecialOffer(Item[id=12]), category=Database, durationInMs=50, finished=true]
TraceEvent[name=getSpecialOffer(Item[id=23]), category=Database, durationInMs=50, finished=true]
TraceEvent[name=getSpecialOffer(Item[id=34]), category=Database, durationInMs=50, finished=true]
}
}
You see, it looks almost identical to the original application.
Idea for further simplification, iteration 2
When reading method Example.getOrder(int orderId) I was wondering why you are calling order.getItems(), looping over it and calling getSpecialOffer(item) inside the loop. In your sample code you do not use the results for anything other than updating the transaction trace object. I am assuming that in your real code you do something with the order and with the special offers in that method.
But just in case you really do not need those calls inside that method, I suggest
you factor the calls out right into the aspect, getting rid of the TransactionTrace annotation parameters String addStat() and String incrementStat().
The Example code would get even simpler and
the annotation #TransactionTrace(message = "", addStat = "number of items") in class would go away, too.
I am leaving this refactoring to you if you think it makes sense.
I need to map two classes with Orika, OrikaA and OrikaB. OrikaB class have a constructor with one parameter that doesn't exist in OrikaA class. To map OrikaA to OrikaB I used a custom ObjectFactory:
public class OrikaA {
String fieldA;
public OrikaA(String fieldA) {
this.fieldA = fieldA;
}
public String getFieldA() {
return fieldA;
}
}
public class OrikaB {
String fieldA;
String fieldB;
public OrikaB(String fieldA, String fieldB) {
this.fieldA = fieldA;
this.fieldB = fieldB;
}
public String getFieldA() {
return fieldA;
}
public String getFieldB() {
return fieldB;
}
}
#Test
public void orikaObjectFactoryTest() {
mapperFactory = new DefaultMapperFactory.Builder().build();
mapperFactory.classMap(OrikaA.class, OrikaB.class).register();
Type<OrikaA> orikaAType = TypeFactory.valueOf(OrikaA.class);
Type<OrikaB> orikaBType = TypeFactory.valueOf(OrikaB.class);
ObjectFactory<OrikaA> objectFactoryOrikaA = (Object o, MappingContext mappingContext) -> {
OrikaB source = (OrikaB) o;
return new OrikaA(source.getFieldA());
};
ObjectFactory<OrikaB> objectFactoryOrikaB = (Object o, MappingContext mappingContext) -> {
OrikaA source = (OrikaA) o;
return new OrikaB(source.getFieldA(), null);
};
mapperFactory.registerObjectFactory(objectFactoryOrikaA, orikaAType, orikaBType);
mapperFactory.registerObjectFactory(objectFactoryOrikaB, orikaBType, orikaAType);
MapperFacade mapper = mapperFactory.getMapperFacade();
OrikaA orikaA = new OrikaA("text");
OrikaB orikaB = mapper.map(orikaA, OrikaB.class);
assertEquals(orikaA.getFieldA(), orikaB.getFieldA());
assertNull(orikaB.getFieldB());
}
Is there another way to accomplish the same task without using the custom ObjectFactory?
Jackson's Polymorphic Serialize/Deserialize capability is really cool, or would be if I could figure out how to apply it to my problem at hand. There is a pretty good article at http://programmerbruce.blogspot.com/2011/05/deserialize-json-with-jackson-into.html that I have been unable to adapt to my simplified problem.
In a nutshell, I am able to get Jackson 2.1.2 to Serialize a class hierarchy into JSON string with type information. I am unable, however, to get Jackson 2.1.2 to Deserialize that JSON String back into my class hierarchy. Below is a Unit Test that exposes this issue.
The class hierarchy is simple enough; Base Class with just two direct Subclasses. Further, the JSON output appears to respect my Jackson #JsonTypeInfo and produces a believable string from mapper.writeValueAsString
{"type":"dog","name":"King","breed":"Collie"}
But my call to mapper.readValue( jsonOfKing, Animal.class ) stacktraces with...
FAILED: testJacksonSerializeDeserialize
com.fasterxml.jackson.databind.JsonMappingException: No suitable constructor found for type [simple type, class org.rekdev.fasterjacksonwtf.PolymorphismTests$Dog]: can not instantiate from JSON object (need to add/enable type information?)
at [Source: java.io.StringReader#32b3a5a0; line: 1, column: 14]
at com.fasterxml.jackson.databind.JsonMappingException.from(JsonMappingException.java:164)
at com.fasterxml.jackson.databind.deser.BeanDeserializer.deserializeFromObjectUsingNonDefault(BeanDeserializer.java:400)
at com.fasterxml.jackson.databind.deser.BeanDeserializer.deserializeFromObject(BeanDeserializer.java:289)
....
Here's my unit test.
import org.testng.annotations.*;
import static org.testng.Assert.*;
import com.fasterxml.jackson.annotation.*;
import com.fasterxml.jackson.annotation.JsonSubTypes.Type;
import com.fasterxml.jackson.databind.*;
public class PolymorphismTests {
#Test
public void testJacksonSerializeDeserialize() throws Exception {
ObjectMapper mapper = new ObjectMapper();
Animal king = new Dog();
king.name = "King";
( (Dog) king ).breed = "Collie";
String jsonOfKing = mapper.writeValueAsString( king );
// JsonMappingException right here!
Animal actualKing = mapper.readValue( jsonOfKing, Animal.class );
assertEquals( king, actualKing );
}
#JsonTypeInfo( use = JsonTypeInfo.Id.NAME, include = JsonTypeInfo.As.PROPERTY, property = "type" )
#JsonSubTypes( { #Type( value = Cat.class, name = "cat" ), #Type( value = Dog.class, name = "dog" ) } )
abstract class Animal {
public String name;
#Override
public abstract boolean equals( Object obj );
#Override
public abstract int hashCode();
}
class Dog extends Animal {
public String breed;
#Override
public boolean equals( Object obj ) {
if ( this == obj ) {
return true;
}
if ( obj == null ) {
return false;
}
if ( getClass() != obj.getClass() ) {
return false;
}
final Dog that = (Dog) obj;
boolean equals = name.equals( that.name ) && breed.equals( that.breed );
return equals;
}
#Override
public int hashCode() {
int hashCode = name.hashCode() + breed.hashCode();
return hashCode;
}
}
class Cat extends Animal {
public String favoriteToy;
#Override
public boolean equals( Object obj ) {
if ( this == obj ) {
return true;
}
if ( obj == null ) {
return false;
}
if ( getClass() != obj.getClass() ) {
return false;
}
final Cat that = (Cat) obj;
boolean equals = name.equals( that.name ) && favoriteToy.equals( that.favoriteToy );
return equals;
}
#Override
public int hashCode() {
int hashCode = name.hashCode() + favoriteToy.hashCode();
return hashCode;
}
}
}
Why won't the ObjectMapper allow me to readValue process the JSON produced by the ObjectMapper.writeValue()?
Make your inner classes static, like:
static class Dog extends Animal { ... }
otherwise things will not work (since non-static inner classes require so-called "implicit this" argument to refer to an instance of enclosing class).