38577605491

Committed 14 Oct 2026 02:06PM UTC coverage: 84.066% (+4.2%) from 79.825%

Build # 38577605491

Build Type

push

github

Committed by

web-flow

Commit Message

Add search vector field (#108)

* add search vector field

* remove build in stemmer, add preparation for snowball stemmer

* enhance tokenizer, info about default locate

* adjust comments

* update language loading

* add normalizer

* fix issues with contao 5.3

* drop 4.13 support

* make indexing async

* update test data

* apply rector

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* address review comments

- Declare an explicit ESCAPE '!' for the LIKE patterns, so escaped wildcards
  match literally on every platform, and test it against stored rows
- Add the missing imports to the README configuration example
- Complete the locale description of BuildSearchVectorEvent
- Replace Contao's internal LegacyLocaleSwitcher in the indexer test with a
  test double

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* move search vector field documentation to docs

Keep a short example in the README like for the other fields and move the
full documentation to docs/dca_fields/search_vector_field.md.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* add createCondition method

* reject several search vector fields per table

A second registration for the same table silently replaced the first one.
Now loading the DCA throws an exception that names the conflicting fields.
Loading a DCA several times in one process, each time registering the field
once, is still allowed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* address review comments on indexing

- Lock the record while its vector is rebuilt (SELECT ... FOR UPDATE in a
  transaction), so concurrent rebuilds cannot overwrite a newer vector;
  requires Doctrine DBAL 3.8
- Expand Contao 5.7 virtual fields before resolving the configured fields
- Reject invalid --id values in the rebuild command instead of rebuilding
  the whole table
- Replace the deprecated Connection::quoteIdentifier() with the platform's
  quoteSingleIdentifier() and report Doctrine deprecat... (continued)

Coverage Stats

392 of 407 new or added lines in 16 files covered. (96.31%)

1588 of 1889 relevant lines covered (84.07%)

8.9 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

75.0

/src/Event/BuildSearchVectorEvent.php

<?php

namespace HeimrichHannot\UtilsBundle\Event;

use Symfony\Contracts\EventDispatcher\Event;

/**
 * Dispatched before a search vector is tokenized and stored.
 *
 * Use it to add values that are not stored in a column of the record, for example tags or other relations kept in a
 * separate table, or to remove values that should not be searchable.
 */
class BuildSearchVectorEvent extends Event
{
    /**
     * @param array<string, mixed> $row    The raw database row
     * @param list<string>         $values The formatted, not yet tokenized values. Each value stays a separate
     *                                     section of the vector, so a quoted phrase never matches across two values.
     * @param string               $locale The language of the record. The values have been resolved in this language,
     *                                     and the values of the vector are normalized and stemmed in it as well.
     */
    public function __construct(
        public readonly string $table,
        public readonly array $row,
        private array $values,
        public readonly string $locale,
    ) {}

    /**
     * @return list<string>
     */
    public function getValues(): array
    {
        return $this->values;
    }

    /**
     * @param list<string> $values
     */
    public function setValues(array $values): void
    {
        $this->values = \array_values($values);
    }

    public function addValue(string $value): void
    {
        $this->values[] = $value;
    }
}

1	<?php
2
3	namespace HeimrichHannot\UtilsBundle\Event;
4
5	use Symfony\Contracts\EventDispatcher\Event;
6
7	/**
8	* Dispatched before a search vector is tokenized and stored.
9	*
10	* Use it to add values that are not stored in a column of the record, for example tags or other relations kept in a
11	* separate table, or to remove values that should not be searchable.
12	*/
13	class BuildSearchVectorEvent extends Event
14	{
15	/**
16	* @param array<string, mixed> $row The raw database row
17	* @param list<string> $values The formatted, not yet tokenized values. Each value stays a separate
18	* section of the vector, so a quoted phrase never matches across two values.
19	* @param string $locale The language of the record. The values have been resolved in this language,
20	* and the values of the vector are normalized and stemmed in it as well.
21	*/
22	public function __construct(
23	public readonly string $table,
24	public readonly array $row,
25	private array $values,
26	public readonly string $locale,
27	) {}	16 ✔
28
29	/**
30	* @return list<string>
31	*/
32	public function getValues(): array
33	{
34	return $this->values;	16 ✔
35	}
36
37	/**
38	* @param list<string> $values
39	*/
40	public function setValues(array $values): void
41	{
NEW 42	$this->values = \array_values($values);	×
43	}
44
45	public function addValue(string $value): void
46	{
47	$this->values[] = $value;	1 ✔
48	}
49	}

heimrichhannot / contao-utils-bundle / 38577605491

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous