Multi-words terms with accents not handled by thesaurus

[rbayet]

If a multi-word synonym or expansion rule contains non ASCII US characters, indexing the thesaurus fails with the error

{"error":{"root_cause":[{"type":"illegal_argument_exception","reason":"failed to build synonyms"}],"type":"illegal_argument_exception","reason":"failed to build synonyms","caused_by":{"type":"parse_exception","reason":"Invalid synonym rule at line 12","caused_by":{"type":"illegal_argument_exception","reason":"term: [multi word synonym] analyzed to a token (multi_word) with position increment != 1 (got: 0)"}}},"status":400}

Exemple:

{"error":{"root_cause":[{"type":"illegal_argument_exception","reason":"failed to build synonyms"}],"type":"illegal_argument_exception","reason":"failed to build synonyms","caused_by":{"type":"parse_exception","reason":"Invalid synonym rule at line 12","caused_by":{"type":"illegal_argument_exception","reason":"term: pâte à modeler analyzed to a token (pâte_à) with position increment != 1 (got: 0)"}}},"status":400}

Possible linked issue : elastic/elasticsearch#27481
Possible way to avoid the error (but with the loss of the invalid synonym rule) : elastic/elasticsearch@88c270d

Preconditions

Magento Version : 2.2.9, EE 2.3.x

ElasticSuite Version : 2.6.9, 2.8.x-dev

Environment : ElasticSearch 6.6 (and 6.8.2)

Third party modules : N/A

Steps to reproduce

1. Create a new Thesaurus, active on all stores
2. Add a reference term "fimo"
3. Add the expansion terms "pâte à modeler"
4. Reindex manually the Thesaurus (CLI) or reindex all

Expected result

1. The Thesaurus index should be rebuilt

Actual result

1. Thesaurus reindexing fails with the error mentionned above

How to fix

In \Smile\ElasticsuiteThesaurus\Model\Indexer\IndexHandler::prepareSynonymFilterData, which alters multi-word synonyms by replacing spaces by a _, add the u flag in order to support all UTF-8 characters and not only ASCII US :

    /**
     * Prepare the thesaurus data to be saved.
     * Spaces are replaced with "_" into multiwords expression (ex foo bar => foo_bar).
     *
     * @param string[] $rows Original thesaurus text rows.
     *
     * @return string[]
     */
    private function prepareSynonymFilterData($rows)
    {
        $rowMaper = function ($row) {
            return preg_replace('/([\w])[\s-](?=[\w])/u', '\1_', $row);
        };

        return array_map($rowMaper, $rows);
    }