User Guide of Magento 2 Elasticsearch Module

After the successful installation of the Magento 2 Elastic Search extension on your Magento 2 store.

The admin will navigate through Stores > Configuration > Catalog > Catalog search > Search engine and then select Webkul Elasticsearch.

Magento 2 Elasticsearch General Settings

Once the search engine got selected as Webkul Elasticsearch, the admin will navigate through Stores > Configuration > Webkul > Elastic Search Setting > General Settings.

Here the Admin has to enter/Select:-

1. Add the elastic server hostname in the Host field.
2. Port Number:- To configure the elastic server enter the port number.
3. Index Prefix:- We are using an elastic search server on multiple websites. If we use an index prefix on websites, then at the time of re-indexing, that particular website data will not affect other websites.
4. Check Connection Status:- Check the current status of the elastic server. 
5. Reset Connection:- The connection will be reset if you choose this option, and all data will be deleted from the current elastic server. The administrator must manually reindex it by going to Elastic search > Index management.
   

Search Settings

Enable Smart Search Mode: Set this to Yes and Webkul applies optimized query algorithms with safe tokenization while hiding the manual query DSL overrides. Set it to No to set the multi-match type, operator, and minimum match yourself.

Show Out-of-Stock Products Separately: Set this to Yes to list sold-out products in their own section below the in-stock results on the catalog search page.

Allow Spell Correction in Search: Pick the fuzziness level for search, from Level 1 upward. Fuzziness lets the engine auto-correct spelling mistakes, so a wrong spelling still returns relevant products.

Select Search Filters: Choose which token filters apply to every search query. Pick from Elision Filter, Lowercase Filter, Synonym Filter, and Stop Word Filter, and select as many as you need.

Enter Comma Separated List Elisions: Enter a comma separated list of elision articles, such as “l, d, t”. This works only when you select the Elision Filter and strips the leading articles from the term.

Use Stop Words Filter: Set this to Yes to drop common words from the search term, so a query like “shoes for men” matches on “shoes” and “men” without stopping on “for”.

Select Language Stemmer: Choose a stemmer that matches your store view language, from the 28 available including English, Arabic, German, French, Hindi, Italian, Norwegian, Portuguese, and Spanish.

Enter Comma Separated Words to Exclude from Stemming: Enter a comma separated list of words to skip stemming, such as “cats, catlike”. They keep their exact form, which helps with brand names and technical terms.

Select Search Filters: Choose which character filters apply to every search query. Pick from Html Strip Char Filter, Mapping Filter, and Pattern Replace Filter to clean up unwanted input.

Enter Comma Separated List Of Filter Mappings: Enter a comma separated list of mappings, such as ٠ => 0, ١ => 1, ٢ => 2, ٣ => 3. Each key swaps with its value, and this works only when you select the Mapping Filter.

Multi-Match Query Configuration

These fields appear when Enable Smart Search Mode is set to No.

Select Frontend Search Type: Choose how the store searches on the frontend. Multi Match Query searches across the fields you select below, while Simple Match Query looks at the product name and SKU only.

Select Fields For Multi Search: Choose the catalog attributes the multi-search evaluates. The default set is sku, description, short_description, url_key, and name, and a product appears when the search term matches any of them.

Select Multi Match Type: Choose the Elasticsearch scoring algorithm for the multi-search. Pick Best Fields, Most Fields, Cross Fields, Phrase, or Phrase Prefix, depending on how closely a product must match.

Select Operator For Multi Search: Choose AND when every term the customer typed must match, or OR when any single term is enough to include a product. Default: OR.

Minimum Should Match: Enter the percentage of query terms a product must match, for example 75%, to reduce low-quality results. A higher value returns fewer but more accurate matches. Default: 75.

Simple Match Query Configuration

Select Frontend Search Type: Simple Match Query restricts search to the product name and SKU only, ignoring the description and short description. It suits B2B portals and parts catalogs where buyers search by SKU or exact part names.

Allow Spell Correction in Search: Pick the fuzziness edit distance the engine allows. Choose Level 0 for exact matches, Level 1 for one edit, Level 2 for two edits, or Auto, so a wrong spelling still returns relevant products.

Select Search Filters: Choose the token filters applied during analysis. Pick from Synonym Filter, Elision Filter, Lowercase Filter, and Stop Word Filter, and you can select more than one at a time.

Enter Comma Separated List Elisions: Enter a comma separated list of elision prefixes to strip, such as “l, d, m, t, s, n, qu” for French or Italian. This works only when you select the Elision Filter.

Use Stop Words Filter: Set this to Yes to remove high-frequency noise words such as and, the, and to from the search term before the query runs.

Enter Comma Separated List Of Stop Words: Enter a comma separated list of stop words to use. If you leave it empty, the module falls back to the Elasticsearch standard default dictionary.

Select Language Stemmer: Choose a stemmer for each store view locale to reduce words to their base root, so a search for running also matches run. Pick the one that matches your store language.

Enter Comma Separated Words to Exclude from Stemming: Enter a comma separated list of words or brand names to preserve without root reduction, such as “shoes, skis”, so they stay searchable in their exact form.

Select Search Filters: Choose the character filters applied before tokenization. Pick from HTML Strip Char Filter, Mapping Filter, and Pattern Replace Filter to clean up unwanted characters in the search term.

Enter Comma Separated List Of Filter Mappings: Enter key value pairs in key => value format to standardize characters, such as mapping Eastern Arabic numerals. This works only when you select the Mapping Filter.

Enter Pattern: Enter a regular expression to identify recurring noise or formatting in product text, such as (\d+)-(\d+) for hyphenated part numbers. This works only when you select the Pattern Replace Filter.

Enter Pattern To Replace With Filter Pattern: Enter the replacement string substituted for the pattern matches, such as $1$2 to unify hyphenated SKUs so 123-456 matches 123456. Leave it blank to strip the matches instead.

Note On Regular Expressions: A badly written regular expression could run very slowly or even throw a StackOverflowError and cause the node it runs on to exit suddenly, so test the pattern on a staging server first.

Search Suggestion in Magento 2 Elastic Search Extension

In search suggestions, with the help of keywords, customers can search categories, pages, or products.

Magento 2 Elasticsearch Did You mean

When a customer enters the wrong product name, it will recommend all nearby products with the same product name to the customers. 

Shown in the below screenshot.

Spell Correction in Search

The Elastic Search Engine will automatically correct keywords when they wrote incorrectly by the customer if the spell check is set to Level 1 or Level 2 by the admin.

Select Search Filters

There are four types of search filters to modify the search term provided by the customer. The admin can select the token filters.

1. Synonym Filter

The admin has to navigate through Marketing > Search Synonyms.

Admin can create a new synonym group from here. For example, the admin has added synonyms for jeans and pants.

Now, the search engine will display all the results connected to any of these keywords if the customer searches for one of them.

2. Elision Filter- An elisions-removing token filter. For example, “L’avion” (the plane) will be parsed as “avion” (plane).

3. Lowercase Filter– A lowercase token filter that converts token text to lowercase.

4. Stop Word Filter– Stop words are removed from the search query. Examples of stop words are and, is, and to.

To learn more about token filters, please go through the following Link:

https://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-tokenfilters.html

Select Language Stemmer

According to the language of the Magento 2(Adobe Commerce) store, the admin must choose the appropriate Language Stemmer. 

For example, if the store language is set as English, the customer can only search in English.

• Select Language Stemmer: Choose the language for stemming words.
• Enter comma separated words to exclude from stemming: List words to keep unchanged during stemming.

Select Character Search Filters in Magento 2 Elasticsearch

There are three types of character filters. That is used to remove unwanted characters from the search input.

1. HTML strip char filter– The HTML strip character filter strips off the HTML parts.
2. Mapping Filter– The mapping filter swaps out the mapping keys with values. So, it will eliminate any character from the mapping in the Magento 2 Elasticsearch module.
3. Pattern Replace Filter– The Pattern Replace Filter functions similarly to the Mapping Filter. The main difference is that it replaces patterns instead of mapping them. To replace characters, it makes use of a regular expression
4. Enter Comma Separated List Of Filter Mappings: List multiple filter rules, separated by commas.
5. Enter Pattern: Define the text or pattern to search for.
6. Enter Pattern To Replace With Filter Pattern: Define the text or pattern to replace the found pattern.

To learn more about the character filters, please go through the following URL:

https://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-charfilters.html