Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance. Project price only 1 $
You can buy this project and download/modify it how often you want.
/*
* Copyright 2021 Adaptive Financial Consulting Ltd.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* https://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package uk.co.real_logic.artio.dictionary.generation;
import org.agrona.collections.Int2ObjectHashMap;
import org.agrona.collections.MutableInteger;
import uk.co.real_logic.artio.dictionary.DictionaryParser;
import uk.co.real_logic.artio.dictionary.ir.Dictionary;
import uk.co.real_logic.artio.dictionary.ir.*;
import uk.co.real_logic.artio.dictionary.ir.Field.Value;
import java.util.*;
import java.util.function.BiFunction;
import java.util.function.BiPredicate;
import java.util.function.Function;
import java.util.stream.Collectors;
import java.util.stream.LongStream;
import java.util.stream.Stream;
import static java.util.stream.Collectors.*;
import static java.util.stream.Stream.concat;
class CodecSharer
{
// Used in sharedNameToField in order to denote a name that has a clash
private static final Field CLASH_SENTINEL = new Field(-1, "SHARING_CLASH", Field.Type.INT);
private static final Comparator> ENUM_NAME_ORDER =
Comparator.>comparingLong(Map.Entry::getValue) // Dictonary Count
.thenComparingInt(e -> e.getKey().length()); // Length of name
private final List inputDictionaries;
private final Map sharedNameToField = new HashMap<>();
private final Map widenedFields = new HashMap<>();
private final Map sharedIdToGroup = new HashMap<>();
private final Set commonGroupIds = new HashSet<>();
private final Map sharedNameToComponent = new HashMap<>();
private final Set commonAnyFields = new HashSet<>();
CodecSharer(final List inputDictionaries)
{
this.inputDictionaries = inputDictionaries;
}
public void share()
{
findSharedFields();
findSharedGroups();
findSharedComponents();
findCommonAnyFields();
final Map nameToSharedMessage = findSharedMessages();
findComponentsWithSharedFieldsNotInAllDictionaries(nameToSharedMessage);
final Component header = findSharedComponent(Dictionary::header);
final Component trailer = findSharedComponent(Dictionary::trailer);
final String specType = DictionaryParser.DEFAULT_SPEC_TYPE;
final int majorVersion = 0;
final int minorVersion = 0;
final List messages = new ArrayList<>(nameToSharedMessage.values());
final Dictionary sharedDictionary = new Dictionary(
messages,
sharedNameToField,
sharedNameToComponent,
header,
trailer,
specType,
majorVersion,
minorVersion);
updateRequiredEntries(sharedDictionary);
sharedDictionary.shared(true);
inputDictionaries.forEach(dict -> connectToSharedDictionary(sharedDictionary, dict));
inputDictionaries.add(sharedDictionary);
}
private void updateRequiredEntries(final Dictionary sharedDictionary)
{
updateRequiredEntriesInAggregates(sharedDictionary.components().values());
updateRequiredEntriesInAggregates(sharedDictionary.messages());
}
private void updateRequiredEntriesInAggregates(final Collection aggregates)
{
aggregates.forEach(sharedComponent ->
{
sharedComponent.allFieldsIncludingComponents().forEach(entry ->
{
final boolean required = entry.required();
entry.sharedChildEntries().forEach(childEntry -> childEntry.required(required));
});
});
}
private void findSharedGroups()
{
findSharedAggregates(
sharedIdToGroup,
commonGroupIds,
dict -> concat(grpPairs(dict.messages()), grpPairs(dict.components().values())).collect(toList()),
(comp, sharedGroup) -> true,
this::copyOf,
this::sharedGroupId);
}
interface GetName extends Function, String>
{
}
interface CopyOf extends BiFunction, T, T>
{
}
private String sharedGroupId(final AggPath pair)
{
return pair.parents().stream().map(Aggregate::name).collect(joining(".")) + "." + pair.agg().name();
}
private Stream> grpPairs(final Collection aggregates)
{
return aggregates.stream().flatMap(agg -> grpPairs(agg, Collections.emptyList()));
}
// Groups within this aggregate and groups within groups, but not groups within components within this aggregate
public Stream> grpPairs(final Aggregate aggregate, final List prefix)
{
final List parents = path(prefix, aggregate);
return aggregate.entriesWith(ele -> ele instanceof Group)
.flatMap(e ->
{
final Group group = (Group)e.element();
final Stream> pair = Stream.of(new AggPath<>(parents, group));
return concat(pair, grpPairs(group, parents));
});
}
private List path(final List prefix, final Aggregate aggregate)
{
final List parents = new ArrayList<>(prefix);
parents.add(aggregate);
return parents;
}
private void findSharedComponents()
{
// find components in all dictionaries
findSharedAggregates(
sharedNameToComponent,
null,
dict -> addFakeParents(dict.components().values()),
(comp, sharedComp) -> true,
this::copyOf,
pair -> pair.agg().name());
}
// Aim is to maximise sharing between dictionaries by
private void findComponentsWithSharedFieldsNotInAllDictionaries(final Map nameToSharedMessage)
{
// Shared messages within dictionaries
final Map> sharedMessageNameToMessages = inputDictionaries
.stream()
.flatMap(dictionary -> dictionary
.messages()
.stream()
.filter(msg -> nameToSharedMessage.containsKey(msg.name())))
.collect(groupingBy(Message::name));
// components not in all dictionaries
final Map> otherNameToComponents = inputDictionaries
.stream()
.flatMap(dictionary -> dictionary
.components()
.entrySet()
.stream()
.filter(e -> !sharedNameToComponent.containsKey(e.getKey())))
.map(Map.Entry::getValue)
.collect(groupingBy(Component::name));
otherNameToComponents.forEach((componentName, components) ->
{
final Set sharedFieldNames = findFieldsInAllInstancesOfComponent(components);
if (sharedFieldNames.isEmpty())
{
return;
}
final Int2ObjectHashMap indexToSynthesizedComponent = new Int2ObjectHashMap<>();
// find shared messages that implement the component in some but not all cases
sharedMessageNameToMessages.forEach((msgName, messages) ->
{
final Int2ObjectHashMap messagesWithoutComponent = new Int2ObjectHashMap<>();
final int withComponentCount = findMessagesWithoutComponent(
componentName, messages, messagesWithoutComponent);
// some but not all
if (withComponentCount > 0 && withComponentCount != messages.size())
{
for (final Map.Entry entry : messagesWithoutComponent.entrySet())
{
final Message message = entry.getValue();
// check it contains all the common fields
final List componentFieldEntries = message
.fieldEntries()
.filter(e -> sharedFieldNames.contains(e.name()))
.collect(toList());
if (sharedFieldNames.size() == componentFieldEntries.size())
{
final int index = entry.getKey();
final Component synthesizedComponent = lookupSynthesizedComponent(
componentName, indexToSynthesizedComponent, componentFieldEntries, index);
final List entries = message.entries();
entries.removeAll(componentFieldEntries);
entries.add(Entry.optional(synthesizedComponent));
}
}
}
});
synthesizeParentComponent(componentName, sharedFieldNames, indexToSynthesizedComponent);
});
}
private void synthesizeParentComponent(
final String componentName,
final Set sharedFieldNames,
final Int2ObjectHashMap indexToSynthesizedComponent)
{
if (!indexToSynthesizedComponent.isEmpty())
{
final Component synthesizedParentComponent = new Component(componentName);
final List entries = synthesizedParentComponent.entries();
for (final String fieldName : sharedFieldNames)
{
entries.add(Entry.optional(sharedNameToField.get(fieldName)));
}
entries.sort(Entry.BY_NAME);
final int entrySize = entries.size();
for (int i = 0; i < entrySize; i++)
{
final Entry parentEntry = entries.get(i);
final int index = i;
parentEntry.sharedChildEntries(indexToSynthesizedComponent
.values()
.stream()
.map(comp -> comp.entries().get(index))
.collect(toList()));
}
sharedNameToComponent.put(componentName, synthesizedParentComponent);
}
}
private Set findFieldsInAllInstancesOfComponent(final List components)
{
final Component component = components.get(0);
final Set sharedFieldNames = component
.fieldEntries()
.map(Entry::name)
.filter(sharedNameToField::containsKey)
.collect(toSet());
final int size = components.size();
for (int i = 1; i < size; i++)
{
final Set fieldNames = components.get(i).fieldEntries().map(Entry::name).collect(toSet());
sharedFieldNames.retainAll(fieldNames);
}
return sharedFieldNames;
}
private int findMessagesWithoutComponent(
final String componentName,
final List messages,
final Int2ObjectHashMap messagesWithoutComponent)
{
int withComponentCount = 0;
final int messagesSize = messages.size();
for (int i = 0; i < messagesSize; i++)
{
final Message msg = messages.get(i);
if (msg.hasComponent(componentName))
{
withComponentCount++;
}
else
{
messagesWithoutComponent.put(i, msg);
}
}
return withComponentCount;
}
private Component lookupSynthesizedComponent(
final String componentName,
final Int2ObjectHashMap indexToComponent,
final List componentFieldEntries,
final int index)
{
Component synthesizedComponent = indexToComponent.get(index);
if (synthesizedComponent == null)
{
synthesizedComponent = new Component(componentName);
synthesizedComponent.isInParent(true);
componentFieldEntries.sort(Entry.BY_NAME);
componentFieldEntries.forEach(fieldEntry -> fieldEntry.isInParent(true));
synthesizedComponent.entries().addAll(componentFieldEntries);
indexToComponent.put(index, synthesizedComponent);
inputDictionaries.get(index).components().put(componentName, synthesizedComponent);
}
return synthesizedComponent;
}
private void findCommonAnyFields()
{
final Map counts = new HashMap<>();
for (final Dictionary inputDictionary : inputDictionaries)
{
for (final Message message : inputDictionary.messages())
{
message.anyFieldsEntries().forEach(entry ->
{
final String id = anyFieldsId(Collections.singletonList(message), (AnyFields)entry.element());
counts.computeIfAbsent(id, k -> new MutableInteger()).increment();
});
}
}
counts.values().removeIf(count -> count.get() < inputDictionaries.size());
commonAnyFields.addAll(counts.keySet());
}
private String anyFieldsId(final List parents, final AnyFields anyFields)
{
return parents.stream().map(Aggregate::name).collect(joining(".")) + "." + anyFields.name();
}
private Map findSharedMessages()
{
return findSharedAggregates(
new HashMap<>(),
null,
dict -> addFakeParents(dict.messages()),
(msg, sharedMessage) -> msg.packedType() == sharedMessage.packedType(),
this::copyOf,
pair -> pair.agg().name());
}
private static final class AggPath
{
private final List parents;
private final T agg;
private AggPath(final List parents, final T agg)
{
this.parents = parents;
this.agg = agg;
}
public List parents()
{
return parents;
}
public T agg()
{
return agg;
}
}
private Collection> addFakeParents(final Collection aggregates)
{
return aggregates
.stream()
.map(agg -> new AggPath(Collections.emptyList(), agg))
.collect(toList());
}
// Groups existing within a Message or Component, so we put the message or component name into the
// name map for the groups, so we can lookup the precise group later on.
// Thus all the getName functions need parent objects
private Map findSharedAggregates(
final Map nameToAggregate,
final Set commonAggregateNamesCopy,
final Function>> dictToAggParentPair,
final BiPredicate check,
final CopyOf copyOf,
final GetName getName)
{
final Set commonAggregateNames = findCommonNames(dict ->
aggregateNames(dictToAggParentPair.apply(dict), getName));
if (commonAggregateNamesCopy != null)
{
commonAggregateNamesCopy.addAll(commonAggregateNames);
}
for (final Dictionary dictionary : inputDictionaries)
{
dictToAggParentPair.apply(dictionary).forEach(e ->
{
final T agg = e.agg();
final String name = getName.apply(e);
if (commonAggregateNames.contains(name))
{
final T sharedAggregate = nameToAggregate.get(name);
if (sharedAggregate == null)
{
final T copy = copyOf.apply(e.parents(), agg);
if (copy != null)
{
nameToAggregate.put(name, copy);
agg.isInParent(true);
}
}
else
{
// merge fields within aggregate
if (!check.test(agg, sharedAggregate))
{
// TODO: if it happens then push the type down into the implementations
System.err.println("Invalid types: ");
System.err.println(agg);
System.err.println(sharedAggregate);
}
else
{
// Still need to merge aggregates even though we have merged fields
// As some fields may be common to different dictionaries but not messages
mergeAggregate(agg, sharedAggregate);
}
}
}
});
}
identifyAggregateEntriesInParent(nameToAggregate, dictToAggParentPair, getName);
return nameToAggregate;
}
private void identifyAggregateEntriesInParent(
final Map nameToAggregate,
final Function>> dictToAggParentPair,
final GetName getName)
{
inputDictionaries.forEach(dictionary ->
{
dictToAggParentPair.apply(dictionary).forEach(e ->
{
final T agg = e.agg();
final String name = getName.apply(e);
final Aggregate sharedAggregate = nameToAggregate.get(name);
if (sharedAggregate != null)
{
identifyAggregateEntriesInParent(agg, sharedAggregate);
}
});
});
}
private void identifyAggregateEntriesInParent(final Aggregate agg, final Aggregate sharedAgg)
{
sharedAgg.entries().forEach(sharedEntry ->
{
agg.entries().forEach(entry ->
{
if (entry.name().equals(sharedEntry.name()))
{
entry.isInParent(true);
}
});
});
}
private void connectToSharedDictionary(final Dictionary sharedDictionary, final Dictionary dict)
{
dict.sharedParent(sharedDictionary);
}
private void findSharedFields()
{
final Set commonNonEnumFieldNames = findCommonNonEnumFieldNames();
final Set allEnumFieldNames = allEnumFieldNames();
for (final Dictionary dictionary : inputDictionaries)
{
final Map fields = dictionary.fields();
commonNonEnumFieldNames.forEach(fieldName -> mergeField(fields, fieldName));
allEnumFieldNames.forEach(enumName -> mergeField(fields, enumName));
}
updateAssociatedLengthFields();
updateFieldTypeWidening();
updateSometimesEnumClashes();
formUnionEnums();
}
private void updateAssociatedLengthFields()
{
sharedNameToField.values().forEach(field ->
DictionaryParser.checkAssociatedLengthField(sharedNameToField, field, "CodecSharer"));
}
private void updateFieldTypeWidening()
{
sharedNameToField.values().removeIf(field -> field == CLASH_SENTINEL);
widenedFields.forEach((name, widenedType) ->
{
inputDictionaries.forEach(dict ->
{
final Field field = dict.fields().get(name);
if (field != null)
{
field.type(widenedType);
}
});
});
}
private void updateSometimesEnumClashes()
{
sharedNameToField.values().forEach(sharedField ->
{
if (sharedField.hasSharedSometimesEnumClash())
{
inputDictionaries.forEach(dict ->
{
final Field dictField = dict.fields().get(sharedField.name());
if (dictField != null)
{
dictField.hasSharedSometimesEnumClash(true);
}
});
}
});
}
private void formUnionEnums()
{
sharedNameToField.values().forEach(field ->
{
if (field.isEnum())
{
final List fieldValues = field.values();
// Rename collisions by name
final Map> nameToReprToCount = fieldValues.stream().collect(
groupingBy(Value::description,
groupingBy(Value::representation, Collectors.counting())));
// Make a new unique Value for every input with the name collisions fixed
final List values = nameToReprToCount
.entrySet().stream().flatMap(e ->
{
final String name = e.getKey();
final Map reprToCount = e.getValue();
final String commonRepr = findCommonName(reprToCount);
final Stream commonValues =
LongStream.range(0, reprToCount.get(commonRepr)).mapToObj(i ->
new Value(commonRepr, name));
final HashSet otherReprs = new HashSet<>(reprToCount.keySet());
otherReprs.remove(commonRepr);
final Stream otherValues = otherReprs.stream().flatMap(repr ->
{
final String newName = name + "_" + DictionaryParser.enumDescriptionToJavaName(repr);
return LongStream.range(0, reprToCount.get(repr))
.mapToObj(i -> new Value(repr, newName));
});
return concat(commonValues, otherValues);
})
.collect(toList());
// Add name collisions by representation to Javadoc
final Map> reprToNameToCount = values.stream().collect(
groupingBy(Value::representation,
groupingBy(Value::description, Collectors.counting())));
final List finalValues = reprToNameToCount
.entrySet().stream().map(e ->
{
final String repr = e.getKey();
final Map nameToCount = e.getValue();
final String commonName = findCommonName(nameToCount);
final Value value = new Value(repr, commonName);
if (nameToCount.size() > 1)
{
final Set otherNames = nameToCount.keySet();
otherNames.remove(commonName);
value.alternativeNames(new ArrayList<>(otherNames));
}
return value;
})
.collect(toList());
fieldValues.clear();
fieldValues.addAll(finalValues);
}
});
}
private String findCommonName(final Map nameToCount)
{
return nameToCount
.entrySet()
.stream()
.max(ENUM_NAME_ORDER)
.get()
.getKey();
}
private void mergeField(final Map fields, final String fieldName)
{
sharedNameToField.compute(fieldName, (name, sharedField) ->
{
final Field field = fields.get(name);
if (field == null)
{
return sharedField;
}
field.isInParent(true);
if (sharedField == null)
{
return copyOf(field);
}
else
{
final Field.Type sharedType = sharedField.type();
final Field.Type type = field.type();
final BaseType sharedBaseType = BaseType.from(sharedType);
final BaseType baseType = BaseType.from(type);
if (sharedType != type && sharedBaseType != baseType)
{
final BaseType widenedBaseType = attemptWidenField(sharedBaseType, baseType);
if (widenedBaseType == null)
{
// No need to widen if we later find a clash.
widenedFields.remove(name);
if (sharedField.isEnum())
{
System.err.println("Clash error for enum: " + sharedField);
System.out.println(field);
}
// Remove sharing of previous instances:
inputDictionaries.forEach(dict ->
{
final Field alreadySharedField = dict.fields().get(name);
alreadySharedField.isInParent(false);
});
return CLASH_SENTINEL;
}
else
{
final Field.Type widenedType = BaseType.to(widenedBaseType);
sharedField.type(widenedType);
widenedFields.put(name, widenedType);
}
}
mergeEnumValues(sharedField, field);
return sharedField;
}
});
}
private BaseType attemptWidenField(final BaseType left, final BaseType right)
{
final BaseType widenedType = attemptWidenFieldOrdered(left, right);
if (widenedType != null)
{
return widenedType;
}
return attemptWidenFieldOrdered(right, left);
}
private BaseType attemptWidenFieldOrdered(final BaseType left, final BaseType right)
{
// Not super-happy about widening char + int to String
final boolean leftChar = left == BaseType.CHAR;
final boolean leftInt = left == BaseType.INT;
final boolean leftTimestamp = left == BaseType.TIMESTAMP;
if (((leftChar || leftInt || leftTimestamp) && right == BaseType.STRING) || leftChar && right == BaseType.INT)
{
return BaseType.STRING;
}
if (leftInt && right == BaseType.TIMESTAMP)
{
return BaseType.STRING;
}
// widen timestamp + string to string and int + timestamp to String
return null;
}
private void mergeEnumValues(final Field sharedField, final Field field)
{
final List sharedValues = sharedField.values();
final List values = field.values();
final boolean sharedIsEnum = sharedValues.isEmpty();
final boolean isEnum = values.isEmpty();
if (sharedIsEnum != isEnum)
{
sharedField.hasSharedSometimesEnumClash(true);
}
sharedValues.addAll(values);
}
private Set allEnumFieldNames()
{
return inputDictionaries
.stream()
.flatMap(dict -> fieldNames(dict, true))
.collect(Collectors.toSet());
}
private Field copyOf(final Field field)
{
final Field newField = new Field(field.number(), field.name(), field.type());
newField.values().addAll(field.values());
return newField;
}
private Component findSharedComponent(final Function getter)
{
final Dictionary firstDictionary = inputDictionaries.get(0);
final Component sharedComponent = copyOf(Collections.emptyList(), getter.apply(firstDictionary));
inputDictionaries.forEach(dict ->
{
final Component component = getter.apply(dict);
mergeAggregate(component, sharedComponent);
});
inputDictionaries.forEach(dict ->
{
final Component component = getter.apply(dict);
identifyAggregateEntriesInParent(component, sharedComponent);
});
return sharedComponent;
}
private void mergeAggregate(final Aggregate aggregate, final Aggregate sharedAggregate)
{
aggregate.isInParent(true);
final Map nameToEntry = nameToEntry(aggregate.entries());
final Iterator it = sharedAggregate.entries().iterator();
while (it.hasNext())
{
final Entry sharedEntry = it.next();
if (sharedEntry.element() == null)
{
it.remove();
continue;
}
final Entry entry = nameToEntry.get(sharedEntry.name());
if (entry == null)
{
it.remove();
}
else
{
// Only required if all are required
sharedEntry.required(sharedEntry.required() && entry.required());
sharedEntry.sharedChildEntries().add(entry);
}
}
}
private Set findCommonNonEnumFieldNames()
{
return findCommonNames(dictionary -> fieldNames(dictionary, false).collect(Collectors.toSet()));
}
private Set findCommonNames(final Function> getAllNames)
{
final Set messageNames = new HashSet<>();
inputDictionaries.forEach(dict ->
{
final Set namesInDictionary = getAllNames.apply(dict);
if (messageNames.isEmpty())
{
messageNames.addAll(namesInDictionary);
}
else
{
messageNames.retainAll(namesInDictionary);
}
});
return messageNames;
}
private Set aggregateNames(
final Collection> aggregates,
final GetName getName)
{
return aggregates.stream().map(getName).collect(Collectors.toSet());
}
private Stream fieldNames(final Dictionary dictionary, final boolean isEnum)
{
return dictionary.fields().entrySet()
.stream()
.filter(e -> e.getValue().isEnum() == isEnum)
.map(Map.Entry::getKey);
}
private Map nameToEntry(final List entries)
{
return entries.stream().collect(Collectors.toMap(Entry::name, x -> x));
}
// pre: shared fields calculated
private Entry copyOf(final List parents, final Entry entry)
{
Entry.Element element = entry.element();
// Remap fields to calculated shared values
if (element instanceof Field)
{
final Field field = (Field)element;
final String name = field.name();
element = sharedNameToField.get(name);
if (element == null)
{
return null;
}
}
else if (element instanceof Component)
{
final Component component = (Component)element;
final String name = component.name();
element = sharedNameToComponent.get(name);
if (element == null)
{
// Component in message but not shared
element = copyOf(parents, component);
}
}
else if (element instanceof Group)
{
final Group group = (Group)element;
final String id = sharedGroupId(new AggPath<>(parents, group));
// Nested Groups need to update this shared id map in order to copy the outer element of a nested group
Group sharedGroup = sharedIdToGroup.get(id);
if (sharedGroup == null)
{
if (!commonGroupIds.contains(id))
{
// Just remove the inner nested group as it isn't in the shared dictionary anyway
return null;
}
sharedGroup = copyOf(parents, group);
sharedIdToGroup.put(id, sharedGroup);
}
element = sharedGroup;
}
else if (element instanceof AnyFields)
{
final String id = anyFieldsId(parents, (AnyFields)element);
if (!commonAnyFields.contains(id))
{
return null;
}
}
else
{
throw new IllegalArgumentException("Unknown element type: " + element);
}
final Entry newEntry = new Entry(entry.required(), element);
final ArrayList sharedChildEntries = new ArrayList<>();
newEntry.sharedChildEntries(sharedChildEntries);
sharedChildEntries.add(entry);
return newEntry;
}
private Component copyOf(final List prefix, final Component component)
{
final Component newComponent = new Component(component.name());
copyTo(prefix, component, newComponent);
return newComponent;
}
private Group copyOf(final List prefix, final Group group)
{
final List parents = path(prefix, group);
final Entry numberField = group.numberField();
final Entry copiedNumberField = copyOf(parents, numberField);
// if number field isn't shared then the group won't be
if (copiedNumberField == null)
{
return null;
}
final Group newGroup = new Group(group.name(), copiedNumberField);
copyTo(prefix, group, newGroup);
return newGroup;
}
private Message copyOf(final List prefix, final Message message)
{
final Message newMessage = new Message(message.name(), message.fullType(), message.category());
copyTo(prefix, message, newMessage);
return newMessage;
}
private void copyTo(final List prefix, final Aggregate aggregate, final Aggregate newAggregate)
{
final List parents = path(prefix, aggregate);
for (final Entry entry : aggregate.entries())
{
final Entry newEntry = copyOf(parents, entry);
if (newEntry != null)
{
newAggregate.entries().add(newEntry);
}
}
}
}
