如何在Magento 2中实现免费的搜索栏拼写纠错功能？

Hey John, great question—dealing with misspelled searches from less tech-savvy customers is such a common pain point, and since you can’t use third-party extensions, building this directly into Magento 2 via custom code is the perfect approach. Let’s walk through how to implement this step by step, using the method you referenced:

We’ll modify Magento’s native search logic to add spell checking. The core idea is to compare the user’s misspelled query against your product catalog (names, SKUs, etc.) using a similarity algorithm (like Levenshtein Distance) to find the closest matching valid term, then surface that as a suggestion or auto-redirect to the correct results.

Always use a custom module instead of modifying core files—it keeps your changes upgrade-safe. Here’s how to build it:

Step 1: Set Up the Custom Module Structure #

Create a new module in app/code/YourVendor/SpellCorrection (replace YourVendor with your actual vendor name):

• registration.php:

  <?php
  \Magento\Framework\Component\ComponentRegistrar::register(
      \Magento\Framework\Component\ComponentRegistrar::MODULE,
      'YourVendor_SpellCorrection',
      __DIR__
  );
  

• etc/module.xml:

  <?xml version="1.0"?>
  <config xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="urn:magento:framework:Module/etc/module.xsd">
      <module name="YourVendor_SpellCorrection" setup_version="1.0.0">
          <sequence>
              <module name="Magento_CatalogSearch"/>
              <module name="Magento_Search"/>
          </sequence>
      </module>
  </config>
  

Step 2: Override the Search Query Model #

We’ll override Magento’s native Query model to add our spell-checking logic.

• etc/di.xml:

  <?xml version="1.0"?>
  <config xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="urn:magento:framework:ObjectManager/etc/config.xsd">
      <preference for="Magento\Search\Model\Query" type="YourVendor\SpellCorrection\Model\Query"/>
  </config>
  

• Model/Query.php:
  This is where the magic happens—we’ll add a method to find the closest matching product term using Levenshtein Distance:

  <?php
  namespace YourVendor\SpellCorrection\Model;
  
  use Magento\Search\Model\Query as OriginalQuery;
  use Magento\Catalog\Model\ResourceModel\Product\CollectionFactory;
  use Magento\Framework\App\ResourceConnection;
  
  class Query extends OriginalQuery
  {
      protected $productCollectionFactory;
      protected $resourceConnection;
  
      public function __construct(
          \Magento\Framework\Model\Context $context,
          \Magento\Framework\Registry $registry,
          \Magento\Search\Model\ResourceModel\Query $resource,
          \Magento\Search\Model\ResourceModel\Query\Collection $resourceCollection,
          \Magento\Search\Model\QueryFactory $queryFactory,
          \Magento\Framework\Session\SessionManagerInterface $session,
          \Magento\Framework\App\Config\ScopeConfigInterface $scopeConfig,
          \Magento\Framework\Stdlib\DateTime\DateTime $dateTime,
          CollectionFactory $productCollectionFactory,
          ResourceConnection $resourceConnection,
          array $data = []
      ) {
          $this->productCollectionFactory = $productCollectionFactory;
          $this->resourceConnection = $resourceConnection;
          parent::__construct(
              $context,
              $registry,
              $resource,
              $resourceCollection,
              $queryFactory,
              $session,
              $scopeConfig,
              $dateTime,
              $data
          );
      }
  
      public function getSuggestedQuery()
      {
          $originalQuery = $this->getQueryText();
          if (empty($originalQuery)) {
              return parent::getSuggestedQuery();
          }
  
          // Find the closest matching term from our product catalog
          $suggestedQuery = $this->findClosestMatchingTerm($originalQuery);
  
          if ($suggestedQuery && $suggestedQuery !== $originalQuery) {
              return $suggestedQuery;
          }
  
          // Fall back to Magento's native suggestion if no match is found
          return parent::getSuggestedQuery();
      }
  
      protected function findClosestMatchingTerm($searchTerm)
      {
          $connection = $this->resourceConnection->getConnection();
          $productTableName = $connection->getTableName('catalog_product_entity_varchar');
          $nameAttributeId = $this->getAttributeId('name');
  
          // Fetch all product names from the database
          $select = $connection->select()
              ->from($productTableName, ['value'])
              ->where('attribute_id = ?', $nameAttributeId)
              ->where('value IS NOT NULL');
  
          $productNames = $connection->fetchCol($select);
  
          $closestTerm = null;
          $minDistance = PHP_INT_MAX;
          // Ignore very short search terms to avoid false matches
          $minSearchLength = 3;
  
          if (strlen($searchTerm) < $minSearchLength) {
              return null;
          }
  
          foreach ($productNames as $name) {
              if (strlen($name) < $minSearchLength) {
                  continue;
              }
              // Calculate similarity between search term and product name
              $distance = levenshtein(strtolower($searchTerm), strtolower($name));
              // Adjust these thresholds based on your users' typical spelling errors
              if ($distance < $minDistance && $distance <= 2 && abs(strlen($name) - strlen($searchTerm)) <= 3) {
                  $minDistance = $distance;
                  $closestTerm = $name;
              }
          }
  
          return $closestTerm;
      }
  
      protected function getAttributeId($attributeCode)
      {
          $connection = $this->resourceConnection->getConnection();
          $eavAttributeTable = $connection->getTableName('eav_attribute');
          $select = $connection->select()
              ->from($eavAttributeTable, ['attribute_id'])
              ->where('entity_type_id = ?', 4) // 4 is the entity ID for products
              ->where('attribute_code = ?', $attributeCode);
  
          return $connection->fetchOne($select);
      }
  }
  

Now we need to show the "Did you mean?" message on the search results page or mini search bar.

Edit your theme’s search template (e.g., app/design/frontend/YourTheme/Magento_Search/templates/result.phtml or form.mini.phtml) and add this code:

<?php $query = $block->getQuery(); ?>
<?php if ($suggestedQuery = $query->getSuggestedQuery()): ?>
    <div class="search-spell-suggestion">
        <?= __('Did you mean:') ?> <a href="<?= $block->getSearchUrl($suggestedQuery) ?>"><?= $suggestedQuery ?></a>
    </div>
<?php endif; ?>

• Add Full-Text Indexing: For large catalogs, add a FULLTEXT index to the product name column in MySQL to speed up similarity checks.
• Cache Suggestions: Store computed search term suggestions in Magento’s cache to avoid repeating database queries.
• Expand to More Fields: Modify the findClosestMatchingTerm method to include SKUs, descriptions, or other product attributes.

Enable the module and compile the code:

bin/magento module:enable YourVendor_SpellCorrection
bin/magento setup:upgrade
bin/magento setup:di:compile
bin/magento cache:clean

内容的提问来源于stack exchange，提问作者John