search-it-search

SearchIt – Executing Searches

The main class is FriendsOfRedaxo\SearchIt\SearchIt (namespace since v7; legacy alias search_it still works but is deprecated).

Constructor

use FriendsOfRedaxo\SearchIt\SearchIt;

$search = new SearchIt(
    int|false $clang = false,       // language ID or false for all languages
    bool $loadSettings = true,       // load addon backend settings
    bool $useStopwords = true        // apply German stopword list
);

Basic search

$searchTerm = rex_request('search', 'string', '');
if ($searchTerm !== '') {
    $search = new SearchIt();
    $result = $search->search($searchTerm);
}

Filtering results

// Only specific articles
$search->searchInArticles([1, 5, 12]);

// Only articles in specific categories
$search->searchInCategories([3, 7]);

// Category tree (category + all subcategories)
$search->searchInCategoryTree(3);

// Only specific media pool categories
$search->searchInFileCategories([2, 4]);

// Only specific DB columns
$search->searchInDbColumn('rex_article', 'name');

// Custom WHERE clause on the index table
$search->setWhere("texttype = 'article'");

YRewrite multi-domain: restrict to one domain

$search->searchInCategoryTree(
    rex_yrewrite::getDomainByName('example.com')->getMountId()
);

Search behaviour

// Logical mode: 'and' (all terms must match) or 'or' (any term)
$search->setLogicalMode('and');

// Text mode: 'plain' (plaintext index), 'unmodified' (HTML), 'both'
$search->setTextMode('plain');

// MySQL search mode: 'like' (LIKE %term%) or 'match' (FULLTEXT MATCH)
$search->setSearchMode('like');

// Weight specific words higher
$search->addWhitelist(['important' => 5, 'redaxo' => 3]);

// Blacklist words
$search->setBlacklist(['stopword1', 'stopword2']);

Highlighting

// Surround tags (default: <mark>…</mark>)
$search->setSurroundTags('<strong>', '</strong>');

// Highlight type
$search->setHighlightType('surroundtext');
// Options: 'sentence', 'paragraph', 'surroundtext', 'surroundtextsingle', 'teaser', 'array'

// Max chars for highlighted text
$search->setMaxHighlightedTextChars(200);

Pagination

$page = rex_request('page', 'int', 1);
$perPage = 10;
$search->setLimit(($page - 1) * $perPage, $perPage);
$result = $search->search($searchTerm);

$totalPages = (int) ceil($result['count'] / $perPage);

Sorting

$search->setOrder(['clang' => 'ASC']);

Result array structure

$result = $search->search('term');

$result['count']              // int – total hits (ignoring LIMIT)
$result['hits']               // array – hit objects
$result['keywords']           // array – parsed search terms with weights
$result['searchterm']         // string – original search string
$result['time']               // float – search duration in seconds
$result['simwords']           // array – similar words (if enabled)
$result['simwordsnewsearch']  // string – suggested search with similar words
$result['blacklisted']        // array|false – blacklisted words found in query
$result['hash']               // string – cache hash

Single hit ($result['hits'][$i])

Key	Type	Description
fid	int/string	Foreign ID (article ID, URL hash, DB primary key)
type	string	'article', 'url', 'db_column', 'file'
table	string	Source table (e.g. rex_article)
column	string/null	Source column (only db_column type)
clang	int/null	Language ID
plaintext	string	Indexed plaintext
highlightedtext	string	Text with highlighted search terms
teaser	string	Short teaser text
article_teaser	string	Teaser from article content
values	array/null	Additional indexed column values
filename	string/null	Filename (only file type)
fileext	string/null	File extension (only file type)

Distinguishing hit types

foreach ($result['hits'] as $hit) {
    switch ($hit['type']) {
        case 'article':
            $url = rex_getUrl($hit['fid'], $hit['clang']);
            $name = rex_article::get($hit['fid'], $hit['clang'])?->getName();
            break;

        case 'db_column':
            if ($hit['table'] === rex::getTable('article')) {
                $url = rex_getUrl($hit['fid'], $hit['clang']);
            }
            break;

        case 'file':
            $url = rex_url::media($hit['filename']);
            break;

        case 'url':
            // URL addon hit – reconstruct URL via profile
            break;
    }
}

Similarity search ("Did you mean?")

Requires activation in backend settings. Builds a keyword index from successful searches over time.

$result = $search->search($searchTerm);

if ($result['count'] === 0 && !empty($result['simwordsnewsearch'])) {
    echo 'Did you mean: <a href="?search=' . urlencode($result['simwordsnewsearch']) . '">'
        . rex_escape($result['simwordsnewsearch']) . '</a>';
}

Namespace structure (v7+)

FriendsOfRedaxo\SearchIt\
├── SearchIt                     – main class (search + indexing)
├── Api\Autocomplete             – autocomplete API endpoint
├── Cache\SearchCache            – search result cache
├── Console\ReindexCommand       – console: search_it:reindex
├── Console\ClearCacheCommand    – console: search_it:clearCache
├── Cronjob\Reindex              – cronjob: reindex
├── Cronjob\ClearCache           – cronjob: clear cache
├── EventHandler                 – REDAXO extension point handler
├── Helper\ArticleHelper         – article/category lists
├── Helper\ColognePhonetic       – Cologne phonetic (similarity)
├── Helper\FileHelper            – file/directory scanning
├── Helper\FormBuilder           – backend settings forms
├── Helper\SocketHelper          – HTTP socket for indexing
├── Helper\UrlAddon              – URL addon integration
├── Index\KeywordStore           – keyword storage
├── Pdf\PdfConverter             – PDF-to-text conversion
├── Plaintext\PlaintextConverter – HTML-to-plaintext
├── Search\Highlighter           – frontend keyword highlighter
├── Stats\Statistics             – search statistics

Common pitfalls

• Forgetting rex_request('search', 'string', '') – never use $_GET['search'] directly.
• Calling $search->search() without checking for empty search term – returns all indexed content, slow on large sites.
• Mixing up searchInCategories() (direct children only) and searchInCategoryTree() (recursive) – use the tree variant for section-scoped search.
• Using setSearchMode('match') without a FULLTEXT index on the search_it tables – falls back silently to slow queries.
• Not regenerating the index after adding new searchInDbColumn() sources in the backend.