io.evitadb.api.requestResponse.schema.builder.InternalSchemaBuilderHelper Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of evita_api Show documentation
Show all versions of evita_api Show documentation
Module contains external API of the evitaDB.
/*
*
* _ _ ____ ____
* _____ _(_) |_ __ _| _ \| __ )
* / _ \ \ / / | __/ _` | | | | _ \
* | __/\ V /| | || (_| | |_| | |_) |
* \___| \_/ |_|\__\__,_|____/|____/
*
* Copyright (c) 2023
*
* Licensed under the Business Source License, Version 1.1 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* https://github.com/FgForrest/evitaDB/blob/master/LICENSE
*
* 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 io.evitadb.api.requestResponse.schema.builder;
import io.evitadb.api.exception.AttributeAlreadyPresentInEntitySchemaException;
import io.evitadb.api.exception.InvalidSchemaMutationException;
import io.evitadb.api.exception.SortableAttributeCompoundSchemaException;
import io.evitadb.api.requestResponse.schema.AttributeSchemaContract;
import io.evitadb.api.requestResponse.schema.CatalogSchemaContract;
import io.evitadb.api.requestResponse.schema.CatalogSchemaEditor;
import io.evitadb.api.requestResponse.schema.EntitySchemaContract;
import io.evitadb.api.requestResponse.schema.EntitySchemaEditor;
import io.evitadb.api.requestResponse.schema.NamedSchemaContract;
import io.evitadb.api.requestResponse.schema.SortableAttributeCompoundSchemaContract;
import io.evitadb.api.requestResponse.schema.SortableAttributeCompoundSchemaContract.AttributeElement;
import io.evitadb.api.requestResponse.schema.builder.InternalEntitySchemaBuilder.AttributeNamingConventionConflict;
import io.evitadb.api.requestResponse.schema.mutation.CombinableCatalogSchemaMutation;
import io.evitadb.api.requestResponse.schema.mutation.CombinableEntitySchemaMutation;
import io.evitadb.api.requestResponse.schema.mutation.EntitySchemaMutation;
import io.evitadb.api.requestResponse.schema.mutation.LocalCatalogSchemaMutation;
import io.evitadb.api.requestResponse.schema.mutation.SchemaMutation;
import io.evitadb.utils.ArrayUtils;
import io.evitadb.utils.Assert;
import javax.annotation.Nonnull;
import javax.annotation.Nullable;
import java.lang.reflect.Array;
import java.util.Arrays;
import java.util.Collection;
import java.util.Iterator;
import java.util.LinkedList;
import java.util.List;
import java.util.Map;
import java.util.Objects;
import java.util.function.BiFunction;
import java.util.stream.Stream;
/**
* Interface contains shared logic both for {@link CatalogSchemaEditor} and {@link EntitySchemaEditor}.
*
* @author Jan Novotný ([email protected]), FG Forrest a.s. (c) 2022
*/
public interface InternalSchemaBuilderHelper {
/**
* Method does the real heavy lifting of adding the mutation to the existing set of mutations.
* This method is quite slow - it has addition factorial complexity O(n!+). For each added mutation we need to
* compute combination result with all previous mutations.
*
* @param mutationType just for the Java generics sake, damn!
* @param combiner again a lambda that satisfies the damned Java generics - we just need to call combineWith method
* @param existingMutations set of existing mutations in the pipeline
* @param newMutations array of new mutations we want to add to the pipeline
* @param :) having Java more clever, we wouldn't need this
* @return TRUE if the pipeline was modified and the cached schema needs te "recalculated"
*/
@SafeVarargs
private static MutationImpact addMutations(
@Nonnull Class mutationType,
@Nonnull BiFunction> combiner,
@Nonnull List existingMutations,
@Nonnull T... newMutations
) {
MutationImpact schemaUpdated = MutationImpact.NO_IMPACT;
// go through all new mutations
for (T newMutation : newMutations) {
if (newMutation instanceof CombinableEntitySchemaMutation || newMutation instanceof CombinableCatalogSchemaMutation) {
final List> replacements = new LinkedList<>();
// for each - traverse all existing mutations
@SuppressWarnings("unchecked")
T[] mutationsToExamine = (T[]) Array.newInstance(mutationType, 1);
mutationsToExamine[0] = newMutation;
do {
final T[] mutationsToGoThrough = Arrays.copyOf(mutationsToExamine, mutationsToExamine.length);
mutationsToExamine = null;
boolean discarded = false;
for (int i = 0; i < mutationsToGoThrough.length; i++) {
T examinedMutation = mutationsToGoThrough[i];
final Iterator existingIt = existingMutations.iterator();
int index = -1;
while (existingIt.hasNext()) {
index++;
final T existingMutation = existingIt.next();
// and try to combine them together
final MutationCombinationResult combinationResult = combiner.apply(
existingMutation, examinedMutation
);
if (combinationResult != null) {
discarded = combinationResult.discarded();
// now check the result
if (combinationResult.origin() == null) {
// or we may find out that the new mutation makes previous mutation obsolete
existingIt.remove();
index--;
schemaUpdated = updateMutationImpactInternal(schemaUpdated, MutationImpact.MODIFIED_PREVIOUS);
examinedMutation = null;
} else if (combinationResult.origin() != existingMutation) {
// or we may find out that the new mutation makes previous mutation partially obsolete
replacements.add(new MutationReplacement<>(index, combinationResult.origin()));
examinedMutation = null;
}
// we may find out that the new mutation is not necessary, or partially not necessary
if (ArrayUtils.isEmpty(combinationResult.current())) {
break;
} else if (combinationResult.current().length == 1 && combinationResult.current()[0] == examinedMutation) {
// continue with this mutation
} else {
//noinspection unchecked
mutationsToExamine = ArrayUtils.mergeArrays(
Arrays.copyOfRange(mutationsToGoThrough, i + 1, mutationsToGoThrough.length),
combinationResult.current()
);
break;
}
}
}
if (!discarded) {
// replace all partially obsolete existing mutations outside the loop to avoid ConcurrentModificationException
for (MutationReplacement replacement : replacements) {
existingMutations.set(replacement.index(), replacement.replaceMutation());
schemaUpdated = updateMutationImpactInternal(schemaUpdated, MutationImpact.ADDED);
}
// clear applied replacements
replacements.clear();
// and if the new mutation still applies, append it to the end
if (examinedMutation != null) {
existingMutations.add(examinedMutation);
schemaUpdated = updateMutationImpactInternal(schemaUpdated, MutationImpact.ADDED);
}
}
}
} while (mutationsToExamine != null);
} else {
existingMutations.add(newMutation);
schemaUpdated = updateMutationImpactInternal(schemaUpdated, MutationImpact.ADDED);
}
}
return schemaUpdated;
}
/**
* Method updates the impact of the mutation on the schema but only if the impact is more significant than
* the existing one.
*
* @param existingImpactLevel the existing impact level
* @param newImpactLevel the new impact level
* @return the new impact level if more significant than the existing one, otherwise the existing one
*/
@Nonnull
private static MutationImpact updateMutationImpactInternal(@Nonnull MutationImpact existingImpactLevel, MutationImpact newImpactLevel) {
if (existingImpactLevel.ordinal() < newImpactLevel.ordinal()) {
existingImpactLevel = newImpactLevel;
}
return existingImpactLevel;
}
/**
* This method is quite slow - it has addition factorial complexity O(n!+). For each added mutation we need to
* compute combination result with all previous mutations.
*/
default MutationImpact addMutations(
@Nonnull CatalogSchemaContract currentCatalogSchema,
@Nonnull List existingMutations,
@Nonnull LocalCatalogSchemaMutation... newMutations
) {
return addMutations(
LocalCatalogSchemaMutation.class,
(existingMutation, newMutation) -> ((CombinableCatalogSchemaMutation) newMutation)
.combineWith(currentCatalogSchema, existingMutation),
existingMutations,
newMutations
);
}
/**
* This method is quite slow - it has addition factorial complexity O(n!+). For each added mutation we need to
* compute combination result with all previous mutations.
*/
default MutationImpact addMutations(
@Nonnull CatalogSchemaContract currentCatalogSchema,
@Nonnull EntitySchemaContract currentEntitySchema,
@Nonnull List existingMutations,
@Nonnull EntitySchemaMutation... newMutations
) {
return addMutations(
EntitySchemaMutation.class,
(existingMutation, newMutation) -> ((CombinableEntitySchemaMutation) newMutation)
.combineWith(currentCatalogSchema, currentEntitySchema, existingMutation),
existingMutations,
newMutations
);
}
/**
* Method updates the impact of the mutation on the schema but only if the impact is more significant than
* the existing one.
*
* @param existingImpactLevel the existing impact level
* @param newImpactLevel the new impact level
* @return the new impact level if more significant than the existing one, otherwise the existing one
*/
@Nonnull
default MutationImpact updateMutationImpact(@Nonnull MutationImpact existingImpactLevel, @Nullable MutationImpact newImpactLevel) {
return updateMutationImpactInternal(existingImpactLevel, newImpactLevel);
}
/**
* Method checks that the attribute schema has all its possible variants for all {@link io.evitadb.utils.NamingConvention}
* unique among all other attribute schemas on the same level of the parent schema.
*/
default void checkNamesAreUniqueInAllNamingConventions(
@Nonnull Collection attributeSchemas,
@Nonnull Collection compoundSchemas,
@Nonnull NamedSchemaContract newSchema
) {
Stream.concat(
attributeSchemas
.stream()
.filter(it -> !(Objects.equals(it.getName(), newSchema.getName()) && newSchema instanceof AttributeSchemaContract))
.flatMap(it -> it.getNameVariants()
.entrySet()
.stream()
.filter(nameVariant -> nameVariant.getValue().equals(newSchema.getNameVariant(nameVariant.getKey())))
.map(
nameVariant -> new AttributeNamingConventionConflict(
it, null, nameVariant.getKey(), nameVariant.getValue()
)
)
),
compoundSchemas
.stream()
.filter(it -> !(Objects.equals(it.getName(), newSchema.getName()) && newSchema instanceof SortableAttributeCompoundSchemaContract))
.flatMap(it -> it.getNameVariants()
.entrySet()
.stream()
.filter(nameVariant -> nameVariant.getValue().equals(newSchema.getNameVariant(nameVariant.getKey())))
.map(
nameVariant -> new AttributeNamingConventionConflict(
null, it, nameVariant.getKey(), nameVariant.getValue()
)
)
)
)
.forEach(conflict -> {
if (newSchema instanceof AttributeSchemaContract newAttributeSchema) {
if (conflict.conflictingAttributeSchema() == null) {
throw new AttributeAlreadyPresentInEntitySchemaException(
conflict.conflictingCompoundSchema(),
newAttributeSchema,
conflict.convention(), conflict.conflictingName()
);
} else {
throw new AttributeAlreadyPresentInEntitySchemaException(
conflict.conflictingAttributeSchema(),
newAttributeSchema,
conflict.convention(), conflict.conflictingName()
);
}
} else if (newSchema instanceof SortableAttributeCompoundSchemaContract newCompoundSchema) {
if (conflict.conflictingAttributeSchema() == null) {
throw new AttributeAlreadyPresentInEntitySchemaException(
conflict.conflictingCompoundSchema(),
newCompoundSchema,
conflict.convention(), conflict.conflictingName()
);
} else {
throw new AttributeAlreadyPresentInEntitySchemaException(
conflict.conflictingAttributeSchema(),
newCompoundSchema,
conflict.convention(), conflict.conflictingName()
);
}
} else {
throw new IllegalStateException("Should not be possible");
}
});
}
/**
* Method checks whether there is any sortable attribute compound using attribute with particular name and
* throws {@link SortableAttributeCompoundSchemaException} if it does.
*
* @throws SortableAttributeCompoundSchemaException when there is sortable attribute compound using attribute
*/
default void checkSortableAttributeCompoundsWithoutAttribute(
@Nonnull String attributeName,
@Nonnull Collection sortableAttributeCompounds
) {
final SortableAttributeCompoundSchemaContract conflictingCompounds = sortableAttributeCompounds
.stream()
.filter(
it -> it.getAttributeElements()
.stream()
.anyMatch(attr -> attributeName.equals(attr.attributeName()))
)
.findFirst()
.orElse(null);
Assert.isTrue(
conflictingCompounds == null,
() -> new SortableAttributeCompoundSchemaException(
"The attribute `" + attributeName + "` cannot be removed because there is sortable attribute compound" +
" relying on it! Please, remove the compound first. ",
conflictingCompounds
)
);
}
/**
* Method checks whether the sortable attribute is not an array type. It's not possible to sort entities that would
* provide multiple values to sort by.
*/
default void checkSortableTraits(@Nonnull String attributeName, @Nonnull AttributeSchemaContract attributeSchema) {
if (attributeSchema.isSortable()) {
Assert.isTrue(
!attributeSchema.getType().isArray(),
() -> new InvalidSchemaMutationException(
"Attribute " + attributeName + " is marked as sortable and thus cannot be the array of " +
attributeSchema.getType() + "!"
)
);
}
}
/**
* Method checks whether the sortable attribute is not an array type. It's not possible to sort entities that would
* provide multiple values to sort by.
*/
default void checkSortableTraits(
@Nonnull String compoundSchemaName,
@Nonnull SortableAttributeCompoundSchemaContract compoundSchemaContract,
@Nonnull Map attributeSchemas
) {
for (AttributeElement attributeElement : compoundSchemaContract.getAttributeElements()) {
final AttributeSchemaContract attributeSchema = attributeSchemas.get(attributeElement.attributeName());
if (attributeSchema == null) {
throw new SortableAttributeCompoundSchemaException(
"Attribute `" + attributeElement.attributeName() + "` the sortable attribute compound" +
" `" + compoundSchemaName + "` consists of doesn't exist!",
compoundSchemaContract
);
} else {
Assert.isTrue(
!attributeSchema.getType().isArray(),
() -> new InvalidSchemaMutationException(
"Attribute `" + attributeElement.attributeName() + "` the sortable attribute compound" +
" `" + compoundSchemaName + "` consists of cannot be the array of " +
attributeSchema.getType() + "!"
)
);
}
}
}
/**
* Enum representing the impact of the mutation on the schema.
*/
enum MutationImpact {
/**
* No mutation was added.
*/
NO_IMPACT,
/**
* All or some of the mutations were added.
*/
ADDED,
/**
* Some mutations might have been added, but already existing mutations were removed or replaced.
*/
MODIFIED_PREVIOUS
}
/**
* Internal record enveloping information for replacing existing mutation in the pipeline with different one.
*
* @param index the index where the existing mutation should be exchanged
* @param replaceMutation the mutation that should replace the existing mutation
*/
record MutationReplacement(
int index,
@Nonnull T replaceMutation
) {
}
/**
* Record representing the order for changing the existing mutation pipeline with different contents.
*
* @param discarded if set to TRUE it continues to be combining `current` mutation with other existing mutations,
* but finally it discards it and does not append it to the list of mutations
* @param origin represents the new mutation that should be used instead of existing mutation, NULL means that
* existing mutation should be discarded without any compensation
* @param current represents the new mutation set that should be appended instead the currently added inserted
* mutation, NULL means that the added mutation should be discarded completely, multiple mutation
* set means that the current mutation should be dismantled to a pieces of which only a few should
* be added to the pipeline
*/
record MutationCombinationResult(
boolean discarded,
@Nullable T origin,
@Nullable T... current
) {
@SafeVarargs
public MutationCombinationResult(@Nullable T origin, @Nullable T... current) {
this(false, origin, current == null || current.length == 0 || current[0] == null ? null : current);
}
}
}