zodiac1978 / wp-search-ignore-block-names

Modifies the native search to ignore block editor comments

Ignore HTML and shortcodes in search

Modifies the native search to ignore HTML markup (block editor comments) and shortcodes

Why?

This plugin modifies the native wordpress search feature to ignore HTML markup (like block names) and shortcodes in your content.

With this plugin activated the unexpected search results are fixed which are reported in #3739 and in #56294

Works only for MySQL 8.0.4+ and MariaDB 10.0.5+

Frequently Asked Questions

I don't see any changes.

This plugin has no settings page, it modifies the internal search to ignore HTML and HTML comments in the post markup.

Imagine a development blog using a syntaxhighlighting block. If you search on this blog to find posts with the term "syntaxhighlighting" you will get every post with code and not only those post which are really about syntax highlighting.

What about the alt/title attribute?

At the moment this is not searched for anymore with this plugin. Maybe I can specifically allow this in a future version of the plugin.

I need more features!

This plugin just wants to solve this particular problem without too much overhead. If you need more customizations to your search, then please have a look at Relevanssi or Better Search.

These plugins are building up an index from your content and are therefore solving the underlying issue in another way with much more ways to customize it. And they will be even faster.

Can I help you?

Thanks for asking! Yes, I'm very interested in edge cases and incompatibilities (other plugins or server configurations). Please open a new issue for this.

Thanks

Props for helping to fix this go to espiat, l1nuxjedi, mustafauysal, ravishaheshan, and websupporter!

Changelog

1.2.0

• Fixed broken search for title and excerpt
• Ignore shortcodes additionally to markup
• Rename plugin

1.1.1

• Added back link to error message, if REGEXP_REPLACE is not supported
• Fixed some typos

1.1.0

• Changed approach to remove the markup (now requires MySQL 8.0.4+ or MariaDB 10.0.5+)

1.0.0

• Initial public release

wp-search-ignore-block-names's Issues

Add more solutions the the problem in the FAQ

If there are problems with my approach that are not solvable, people maybe look for alternative solutions. We could provide them in the FAQ:

Possible alternatives:

• Elastic Search
• Relevanssi (https://wordpress.org/plugins/relevanssi/)

Problem with Full Site Editor showing the no results block

Reported upstream:
WordPress/gutenberg#49575
https://core.trac.wordpress.org/ticket/58080

Is REGEXP_REPLACE slow?

Adding a toggle for the internal search in admin area

Looks like an excellent solution. Does this filter backend searches as well? I find that searching for blocks can be quite useful there. Might be worth adding an is_admin()...
Source: Roy Tanck (@roytanck) on Twitter

To allow searching for blocks as well, we could add a toggle to enable/disable searching for block comments.

Search for title and excerpt is broken with this plugin

A simplified Query for a standard search is this

SELECT SQL_CALC_FOUND_ROWS wp_posts.ID 
FROM wp_posts WHERE 1=1 
AND (((wp_posts.post_title LIKE '%wat%') 
OR (wp_posts.post_excerpt LIKE '%wat%') 
OR (wp_posts.post_content LIKE '%wat%')))

This plugin adds a new condition:

$where .= " AND REGEXP_REPLACE({$wpdb->posts}.post_content, '<.+?>', '') LIKE '%{$search_query}%'";

This breaks searching for title and excerpt.

We need to add the two OR statements again to fix this.

Add back link to plugin page after deactivation

Prevent activation if database is not supporting REGEXP_REPLACE

Add HTML markup to the ignored text

We could also ignore HTML markup like the plugin I was following to achieve the goal:
https://wordpress.org/plugins/wp-search-ignore-html-tags/

Support ignoring shortcodes

Trying to do the same for shortcodes:
$where .= " AND REGEXP_REPLACE({$wpdb->posts}.post_content, '\[.+?\]', '') LIKE '%{$search_query}%'";

But although it works in my regex test ...
https://regex101.com/r/lbyiyL/1

... it does not work for the search.

Maybe @LinuxJedi can shed some light for me again :)

Not working as expected

Your solution doesn't work when the search term is in both tags and content. ex: this is a paragraph the post with this content doesn't get captured when searching for "paragraph" because of the AND clause in your query

See: https://core.trac.wordpress.org/ticket/56294#comment:5

Unfortunately I did not find a simple way to achieve this with MySQL below 8.
Since MySQL 8 there is REPLACE_REGEXP which would solve this:
https://stackoverflow.com/questions/986826/how-to-do-a-regular-expression-replace-in-mysql

See: https://core.trac.wordpress.org/ticket/56294#comment:6

I will try to get this working at least with MySQL 8+ (and with the according MariaDB version). If this works, there needs to be some MySQL version checking built in.