IBX-9846: Added search using embeddings

src/bundle/Core/DependencyInjection/Configuration/Parser/Embeddings.php

@@ -0,0 +1,70 @@
+<?php
+
+/**
+ * @copyright Copyright (C) Ibexa AS. All rights reserved.
+ * @license For full copyright and license information view LICENSE file distributed with this source code.
+ */
+declare(strict_types=1);
+
+namespace Ibexa\Bundle\Core\DependencyInjection\Configuration\Parser;
+
+use Ibexa\Bundle\Core\DependencyInjection\Configuration\AbstractParser;
+use Ibexa\Bundle\Core\DependencyInjection\Configuration\SiteAccessAware\ContextualizerInterface;
+use Symfony\Component\Config\Definition\Builder\NodeBuilder;
+
+/**
+ * Configuration parser for embedding models.
+ *
+ * Example configuration:
+ * ```yaml
+ * ibexa:
+ *   system:
+ *      default: # configuration per siteaccess or siteaccess group
+ *          embedding_models:
+ *              name: "text-embedding-3-small"
+ *              dimensions: 1536
+ *              field_suffix: "3small"
+ *              embedding_provider: "ibexa_openai"
+ *          default_embedding_model: text-embedding-ada-002
+ * ```
+ */
+final class Embeddings extends AbstractParser
+{
+    public function addSemanticConfig(NodeBuilder $nodeBuilder): void
+    {
+        $nodeBuilder
+            ->arrayNode('embedding_models')
+                ->normalizeKeys(false)
+                ->info('Defines available embedding models')
+                ->arrayPrototype()
+                    ->children()
+                        ->scalarNode('name')->isRequired()->end()
+                        ->integerNode('dimensions')->isRequired()->end()
+                        ->scalarNode('field_suffix')->isRequired()->end()
+                        ->scalarNode('embedding_provider')->isRequired()->end()
+                    ->end()
+                ->end()
+            ->end()
+            ->scalarNode('default_embedding_model')
+                ->info('Default embedding model identifier')
+                ->defaultValue('text-embedding-ada-002')
+            ->end();
+    }
+
+    /**
+     * @param array<mixed> $config
+     */
+    public function preMap(array $config, ContextualizerInterface $contextualizer): void
+    {
+        $contextualizer->mapConfigArray('embedding_models', $config);
+        $contextualizer->mapSetting('default_embedding_model', $config);
+    }
+
+    /**
+     * @param array<mixed> $scopeSettings
+     */
+    public function mapConfig(array &$scopeSettings, $currentScope, ContextualizerInterface $contextualizer): void
+    {
+        // Nothing to do here.
+    }
+}

src/bundle/Core/IbexaCoreBundle.php

@@ -123,6 +123,7 @@ public function getContainerExtension()
                     new ConfigParser\UrlChecker(),
                     new ConfigParser\TwigVariablesParser(),
                     new ConfigParser\UserContentTypeIdentifier(),
+                    new ConfigParser\Embeddings(),
                 ],
                 [
                     new RepositoryConfigParser\Storage(),

src/bundle/Core/Resources/config/default_settings.yml

@@ -273,3 +273,6 @@ parameters:
         writeFlags: ~
         linkHandling: ~
         permissions: [ ]
+
+    ibexa.site_access.config.default.embedding_models: []
+    ibexa.site_access.config.default.default_embedding_model: 'text-embedding-ada-002'

src/bundle/Core/Resources/config/embeddings.yaml

@@ -0,0 +1,24 @@
+services:
+    _defaults:
+        autowire: true
+        autoconfigure: true
+        public: false
+
+    Ibexa\Contracts\Core\Search\Embedding\EmbeddingProviderRegistryInterface:
+        alias: Ibexa\Core\Search\Embedding\EmbeddingProviderRegistry
+
+    Ibexa\Core\Search\Embedding\EmbeddingProviderRegistry:
+        arguments:
+            $embeddingProviders: !tagged_iterator { tag: 'ibexa.embedding_provider', index_by: 'provider_name' }
+
+    Ibexa\Contracts\Core\Search\Embedding\EmbeddingProviderResolverInterface:
+        alias: Ibexa\Core\Search\Embedding\EmbeddingProviderResolver
+
+    Ibexa\Core\Search\Embedding\EmbeddingProviderResolver: ~
+
+    Ibexa\Contracts\Core\Search\Embedding\EmbeddingConfigurationInterface:
+        alias: Ibexa\Core\Search\Embedding\EmbeddingConfiguration
+
+    Ibexa\Core\Search\Embedding\EmbeddingConfiguration: ~
+
+    Ibexa\Contracts\Core\Search\FieldType\EmbeddingFieldFactory: ~

src/bundle/Core/Resources/config/services.yml

@@ -1,5 +1,6 @@
 imports:
     - { resource: commands.yml }
+    - { resource: embeddings.yaml }
 
 parameters:
     ibexa.site_access.default.name: default

src/contracts/Repository/Values/Content/EmbeddingQuery.php

@@ -0,0 +1,113 @@
+<?php
+
+/**
+ * @copyright Copyright (C) Ibexa AS. All rights reserved.
+ * @license For full copyright and license information view LICENSE file distributed with this source code.
+ */
+declare(strict_types=1);
+
+namespace Ibexa\Contracts\Core\Repository\Values\Content;
+
+use Ibexa\Contracts\Core\Repository\Values\Content\Query\Criterion;
+use Ibexa\Contracts\Core\Repository\Values\Content\Query\Embedding;
+use InvalidArgumentException;
+
+final class EmbeddingQuery extends Query
+{
+    private ?Embedding $embedding = null;
+
+    public function getEmbedding(): ?Embedding
+    {
+        return $this->embedding;
+    }
+
+    public function setEmbedding(?Embedding $embedding): void
+    {
+        $this->embedding = $embedding;
+    }
+
+    public function getFilter(): Criterion
+    {
+        return $this->filter;
+    }
+
+    public function setFilter(Criterion $filter): void
+    {
+        $this->filter = $filter;
+    }
+
+    /**
+     * @return \Ibexa\Contracts\Core\Repository\Values\Content\Query\Aggregation[]
+     */
+    public function getAggregations(): array
+    {
+        return $this->aggregations;
+    }
+
+    /**
+     * @param \Ibexa\Contracts\Core\Repository\Values\Content\Query\Aggregation[] $aggregations
+     */
+    public function setAggregations(array $aggregations): void
+    {
+        $this->aggregations = $aggregations;
+    }
+
+    public function getOffset(): int
+    {
+        return $this->offset;
+    }
+
+    public function setOffset(int $offset): void
+    {
+        $this->offset = $offset;
+    }
+
+    public function getLimit(): int
+    {
+        return $this->limit;
+    }
+
+    public function setLimit(int $limit): void
+    {
+        $this->limit = $limit;
+    }
+
+    public function setPerformCount(bool $performCount): void
+    {
+        $this->performCount = $performCount;
+    }
+
+    public function getPerformCount(): bool
+    {
+        return $this->performCount;
+    }
+
+    public function isValid(): bool
+    {
+        $invalid = [];
+
+        if ($this->query !== null) {
+            $invalid[] = 'query';
+        }
+        if (!empty($this->sortClauses)) {
+            $invalid[] = 'sortClauses';
+        }
+        if (!empty($this->facetBuilders)) {
+            $invalid[] = 'facetBuilders';
+        }
+        if ($this->spellcheck !== null) {
+            $invalid[] = 'spellcheck';
+        }
+
+        if (count($invalid) > 0) {
+            throw new InvalidArgumentException(
+                sprintf(
+                    'EmbeddingQuery did not set [%s].',
+                    implode(', ', $invalid)
+                )
+            );
+        }
+
+        return true;
+    }
+}

src/contracts/Repository/Values/Content/EmbeddingQueryBuilder.php

@@ -0,0 +1,111 @@
+<?php
+
+/**
+ * @copyright Copyright (C) Ibexa AS. All rights reserved.
+ * @license For full copyright and license information view LICENSE file distributed with this source code.
+ */
+declare(strict_types=1);
+
+namespace Ibexa\Contracts\Core\Repository\Values\Content;
+
+use Ibexa\Contracts\Core\Repository\Values\Content\Query\Criterion;
+use Ibexa\Contracts\Core\Repository\Values\Content\Query\Embedding;
+
+final class EmbeddingQueryBuilder
+{
+    private ?Embedding $embedding = null;
+
+    private ?int $limit = null;
+
+    private ?int $offset = null;
+
+    private ?Criterion $filter = null;
+
+    /** @var array<\Ibexa\Contracts\Core\Repository\Values\Content\Query\Aggregation> */
+    private array $aggregations = [];
+
+    private bool $performCount = false;
+
+    private function __construct()
+    {
+    }
+
+    public static function create(): self
+    {
+        return new self();
+    }
+
+    public function withEmbedding(Embedding $embed): self
+    {
+        $this->embedding = $embed;
+
+        return $this;
+    }
+
+    public function setLimit(int $limit): self
+    {
+        $this->limit = $limit;
+
+        return $this;
+    }
+
+    public function setOffset(int $offset): self
+    {
+        $this->offset = $offset;
+
+        return $this;
+    }
+
+    public function setFilter(Criterion $filter): self
+    {
+        $this->filter = $filter;
+
+        return $this;
+    }
+
+    /**
+     * @param array<\Ibexa\Contracts\Core\Repository\Values\Content\Query\Aggregation> $aggregations
+     */
+    public function setAggregations(array $aggregations): self
+    {
+        $this->aggregations = $aggregations;
+
+        return $this;
+    }
+
+    public function setPerformCount(bool $performCount): self
+    {
+        $this->performCount = $performCount;
+
+        return $this;
+    }
+
+    public function build(): EmbeddingQuery
+    {
+        $query = new EmbeddingQuery();
+
+        if ($this->embedding !== null) {
+            $query->setEmbedding($this->embedding);
+        }
+
+        if ($this->limit !== null) {
+            $query->setLimit($this->limit);
+        }
+
+        if ($this->offset !== null) {
+            $query->setOffset($this->offset);
+        }
+
+        if ($this->filter !== null) {
+            $query->setFilter($this->filter);
+        }
+
+        if (!empty($this->aggregations)) {
+            $query->setAggregations($this->aggregations);
+        }
+
+        $query->setPerformCount($this->performCount);
+
+        return $query;
+    }
+}

src/contracts/Repository/Values/Content/Query.php

@@ -14,7 +14,7 @@
 /**
  * This class is used to perform a Content query.
  */
-class Query extends ValueObject
+class Query extends ValueObject implements QueryValidatorInterface
 {
     public const SORT_ASC = 'ascending';
     public const SORT_DESC = 'descending';
@@ -102,6 +102,11 @@ class Query extends ValueObject
      * @var bool
      */
     public $performCount = true;
+
+    public function isValid(): bool
+    {
+        return true;
+    }
 }
 
 class_alias(Query::class, 'eZ\Publish\API\Repository\Values\Content\Query');

src/contracts/Repository/Values/Content/Query/Embedding.php

@@ -0,0 +1,33 @@
+<?php
+
+/**
+ * @copyright Copyright (C) Ibexa AS. All rights reserved.
+ * @license For full copyright and license information view LICENSE file distributed with this source code.
+ */
+declare(strict_types=1);
+
+namespace Ibexa\Contracts\Core\Repository\Values\Content\Query;
+
+use Ibexa\Contracts\Core\Repository\Values\ValueObject;
+
+abstract class Embedding extends ValueObject
+{
+    /** @var float[] */
+    protected array $value;
+
+    /**
+     * @param float[] $value
+     */
+    public function __construct(array $value)
+    {
+        $this->value = $value;
+    }
+
+    /**
+     * @return float[]
+     */
+    public function getValue(): array
+    {
+        return $this->value;
+    }
+}