SearchEngine is a module that creates an index of page contents, and makes it easy to search.
- Install the SearchEngine module.
The module will automatically create an index field during installation, so you should define a custom field (via site config) before installing if you don't want it to be called "search_index". You can change the field name later as well, but you'll have to update the "index_field" option in site config or module settings (in Admin) after renaming the field.
- Add the index field to templates you want to make searchable.
You can add the index field to templates via template settings, field settings (the "Actions" tab), or via the "indexed templates" setting found from the config screen of the SearchEngine module. After the field has been added to a template, every time a page using that template is saved, content from indexable fields will be automatically stored into the (typically hidden) index field.
- Use selectors to query values in site search index or add appropriate render calls to your template file(s).
There are three ways to use SearchEngine, so you'll want to choose the one that fits your workflow best:
- You can query the
search_indexfield (or whatever you decided to name it earlier) just like any other textarea field in ProcessWire. This allows you to build completely customized search feature on your site, while SearchEngine takes care of indexing content. - You can use SearchEngine to find results by calling
$modules->get('SearchEngine')->find($input->get->q). Obviously you can pass any search string to the find method, we're using the "q" GET param here just as an example. - You can let SearchEngine handle the whole search feature by adding a render call to your search page's template file:
$modules->get('SearchEngine')->render(). Therender()method is a shortcut for calling multiple methods behind the scenes, but you can also omit that and instead callrenderStyles(),renderScripts(),renderResults(), andrenderForm()separately.
If you prefer letting SearchEngine handle all the markup for you, the simple approach only involves one method call:
<?php namespace ProcessWire; ?>
<body>
<?= $modules->get('SearchEngine')->render() ?>
</body>
You can pass an array to render(), rendering only specific features. Here's an example of rendering just the form and results list while omitting styles and/or scripts:
<?php namespace ProcessWire; ?>
<body>
<?= $modules->get('SearchEngine')->render(['form', 'results']) ?>
</body>
Finally, here's how you could render all parts separately – this approach involves some additional steps, but provides more control over the rendered output than other methods:
<?php namespace ProcessWire;
$searchEngine = $modules->get('SearchEngine');
...
<head>
<?= $searchEngine->renderStyles() ?>
<?= $searchEngine->renderScripts() ?>
</head>
<body>
<?= $searchEngine->renderForm() ?>
<?= $searchEngine->renderResults() ?>
</body>
SearchEngine supports indexing multilingual content. Once ProcessWire's native multilanguage features are enabled, support for multilanguage indexing can be enabled by converting the index field to FieldtypeTextareaLanguage.
If you render the results list with one of the methods mentioned earlier, SearchEngine will automatically look for the query param defined in its settings – and if this param is found (GET), it will perform a search query for you. If you don't want the module to handle this part for you, you can also perform the query yourself (or just don't render the results list).
You can access the search index field just like any other ProcessWire field with selectors:
if ($q = $sanitizer->selectorValue($input->get->q)) {
// This finds pages matching the query string and returns them as a PageArray:
$results = $pages->find('search_index%=' . $q . ', limit=25');
// Render results and pager with PageArray::render() and PageArray::renderPager():
echo $results->render(); // PageArray::render()
echo $results->renderPager(); // PageArray::renderPager()
// ... or you iterate over the results and render them manually:
echo "<ul>";
foreach ($results as $result) {
echo "<li><a href='{$result->url}'>{$result->title}</a></li>";
}
echo "</ul>";
}Note: while you can use any operator for your selectors, you will likely find the *= and %= operators most useful while querying the index. You can read more about selector operators from ProcessWire's documentation on selectors. By default this module will use the *= operator for built-in features.
Alternatively you can delegate the find operation to the SearchEngine module:
// This performs the query and returns a \SearchEngine\Query object:
$query = $modules->get('SearchEngine')->find($input->get->q);
// This object can be rendered manually, or you can use \SearchEngine\Renderer to render it:
// (First param is an array of custom arguments, second is the Query object.)
echo $modules->get('SearchEngine')->renderResults([], $query);The Renderer class provides support for returning search results as a JSON string, which can be particularly useful for implementing AJAX search features. JSON output is returned by the Renderer::renderResultsJSON() method, and you can customise returned keys and values via results_json_fields configuration option.
The results_json_fields is an array of keys and field names. There are two "special" field name prefixes – template. and parent. – which provide access to template and parent page properties.
Here's an example of how you might use the renderResultsJSON() method in your own code:
// Get the SearchEngine module.
$se = $modules->get('SearchEngine');
// Find pages matching our keyword ("composer"). You could also omit the query and let SearchEngine
// grab the keyword from the configured query GET param (if present).
$query = $se->find('composer');
// Return results as JSON.
$json = $se->renderResultsJSON([
'results_json_fields' => [
'title' => 'title',
'desc' => 'summary',
'url' => 'url',
'parent' => 'parent.title',
'template' => 'template.label',
],
'results_json_options' => JSON_PRETTY_PRINT,
], $query);
var_dump($json);
... which, for an example, might result in following output:
{
"results": [
{
"title": "Git ignore (.gitignore) file",
"desc": "<p>The Wireframe boilerplate site profile includes an opinionated .gitignore file based on Bare Minimum Git project.<\/p>",
"url": "\/docs\/wireframe-boilerplate\/git-ignore-file\/",
"parent": "Wireframe boilerplate",
"template": "Basic page"
},
{
"title": "Directory structure",
"desc": "<p>The directory structure outlined here is loosely based on the <a href=\"http:\/\/framework.zend.com\/manual\/1.12\/en\/project-structure.project.html\" rel=\"nofollow\">recommended project directory structure for Zend Framework 1.x<\/a>. Each component has it's place in the tree, and each directory exists for a reason.<\/p>",
"url": "\/docs\/directory-structure\/",
"parent": "Docs",
"template": "Basic page"
},
{
"title": "Wireframe boilerplate",
"desc": "<p>Wireframe boilerplate is a ProcessWire starter site profile based on the Wireframe output framework.<\/p>",
"url": "\/docs\/wireframe-boilerplate\/",
"parent": "Docs",
"template": "Basic page"
},
{
"title": "ProcessWire Composer Installer",
"desc": "<p>ProcessWire Composer Installer provides <a href=\"https:\/\/getcomposer.org\/doc\/articles\/custom-installers.md\" rel=\"nofollow\">Composer custom installers<\/a> for ProcessWire CMS\/CMF modules and site profiles. While not strictly required by Wireframe, it is an easy way to get started with either Wireframe or the Wireframe Boilerplate site profile.<\/p>",
"url": "\/docs\/processwire-composer-installer\/",
"parent": "Docs",
"template": "Basic page"
}
],
"count": 4,
"total": 4
}
Note that a returned field value (for a specific search result) may be an empty string, but it may also be null. Null return values occur, for an example, if you've specified a field that doesn't exist for that particular page.
If you want to rebuild (recreate) the search index for all pages or pages matching a specific selector, you can do that via the Admin GUI (module configuration screen), or you can perform following request via the API:
// All indexable pages:
$modules->get('SearchEngine')->indexPages();
// Indexable pages matching a selector string:
$modules->get('SearchEngine')->indexPages('template=basic-page');
// Alternatively index just a single page (passing in a Page object):
$modules->get('SearchEngine')->indexPage($page);By default the module will create a search index field called 'search_index' and store values from Page fields title, headline, summary, and body to said index field when a page is saved. You can modify the default behaviour via the Module config screen in the PocessWire Admin, or by defining $config->SearchEngine array in your site config file or other applicable location.
Below you'll find a summary of available options. Note that this is not a comprehensive list – please check the $defaultOptions array near the beginning of the SearchEngine.module.php file for a list with all available options. Some options can be modified via the admin (module config screen), while all options can be defined via code – and if you define an option via site config, it can't be overridden via admin.
You can also pass custom options to certain method calls. Most render-prefixed methods, for an example, accept an array of custom render arguments (see "render_args" in the list below) so that you can optionally override module and site defaults on a case-by-case basis.
$config->SearchEngine = [
// Index field defines the field to store indexed data in. Default value is "search_index".
'index_field' => 'search_index',
// This is an array of all the fields you want to index. Currently there is no way to
// index "all compatible fields", but such an option may be added in the future. (This
// is, in fact, a matter of security: by indexing *all* fields you could inadvertently
// expose sensitive data.)
'indexed_fields' => [
// ...
],
// This is an array of fieldtypes that the module should be considered compatible with.
// You should only modify these values if you're sure that the field in question is
// indeed compatible with the indexing process.
'compatible_fieldtypes' => [
// ...
],
// Currently prefixes contains just the "link" prefix – more may be added in the future.
'prefixes' => [
// The "link" prefix is added to indexed links and enables "link:https://..." queries.
'link' => 'link:',
],
// Find arguments are pretty self-explanatory. These are the defaults and apply only to
// SearchEngine::find(); custom selectors are not affected.
'find_args' => [
'limit' => 20,
'sort' => 'sort',
'operator' => '*=',
'query_param' => 'q',
// The "selector_extra" option is appended to automatically generated selector string.
// You might use this to, for example, return just a subset of all indexed templates.
'selector_extra' => '',
],
// This array is passed directly to the MarkupPagerNav core module, see
// https://processwire.com/docs/front-end/markup-pager-nav/ for more details.
'pager_args' => [
// ...
],
// Render arguments affect the rendered output from the SearchEngine module.
'render_args' => [
// At the moment there's only one "theme" available ("default").
'theme' => 'default',
// By default theme-files (styles and scripts) are served minified; setting
// this option to `false` will serve unmodified, original assets instead.
'minified_resources' => true,
// Various attributes used by the search form and results list when rendered.
'form_action' => './',
'form_id' => 'se-form',
'form_input_id' => 'se-form-input',
'results_summary_id' => 'se-results-summary',
'results_id' => 'se-results',
// Summary of each result (in the search results list) is the value of this field.
'result_summary_field' => 'summary',
// Highlighting basically wraps each instance of the search term in the summary
// text with a predefined tag and class (default element is `<strong></strong>`).
'results_highlight_query' => true,
// These settings define the fields used when search results are rendered as JSON.
'results_json_fields' => [
'title' => 'title',
'desc' => 'summary',
'url' => 'url',
],
// This integer is passed to the json_encode call in renderResultsJSON method.
// See https://www.php.net/json_encode for supported values.
'results_json_options' => 0,
// This defines whether a pager should be rendered at the end of the results list.
'pager' => true,
// Array of classes to use in templates. You can add new classes here as well.
'classes' => [
// ...
],
// Array of strings to use in templates. You can add new classes here as well. Note
// that most of these are `null` in the $defaultOptions array; the reason for this
// is simply that methods cannot be called when declaring class properties, so the
// values are set in SearchEngine::__construct().
'strings' => [
// ...
],
// Template strings. These are used to render markup on the site.
'templates' => [
// ...
],
],
// Currently the requirements array only holds the query min length argument.
'requirements' => [
'query_min_length' => 3,
],
];As of version 0.24.0, SearchEngine has the capability to automatically generate descriptions for search results. You can trigger this behaviour by specifying _auto_desc as the name of the desc key. Note, though, that this comes with some notable caveats:
- Before enabling this feature, you need to be absolutely certain that everything you're indexing can also be publicly displayed. When SearchEngine generates the description automatically, it doesn't know which parts of the index are intended to be displayed publicly, and which parts might've been added to the index just to make them searchable.
- Automatically generated descriptions may, at least for the time being, contain data from multiple fields. This means that such snippets may not be entirely logical (and they very likely won't be complete sentences etc.)
SearchEngine supports the concept of themes. Themes are located in the themes directory, under their own subdirectories (e.g. /themes/theme-name/), and each of them needs to include (at the very least) a config file (/themes/theme-name/config.php). Here's an example of a theme config file:
<?php namespace ProcessWire;
$theme_args = [
'theme_styles' => [
// array of styles, such as:
[
'name' => 'style',
'ext' => 'css',
],
],
'theme_scripts' => [
// array of scripts, such as:
[
'name' => 'script',
'ext' => 'js',
],
],
'render_args' => [
// array of render arguments, such as:
'form_id' => 'my-search-form',
'strings' => [
'results_heading' => __('Custom results heading'),
],
],
'pager_args' => [
// array of pager options
],
];
A theme may include style and script file(s), in which case there should also be a type-specific directory to store them in (i.e. /themes/theme-name/scripts/ or /themes/theme-name/styles/). For each file there should be both the "source" file named in the theme_styles or theme_scripts array (e.g style.css) and a minified version with ".min" in its name (e.g. style.min.css).
Note: modifying files within the module directory is not recommended, since they may get overwritten in an update. You can configure the path for your custom themes via the "Advanced" section of the SearchEngine module config screen.
- ProcessWire >= 3.0.112
- PHP >= 7.1.0
This module can be installed – just like any other ProcessWire module – by downloading or cloning the SearchEngine directory into your /site/modules/ directory. Alternatively you can install SearchEngine with Composer by running composer require teppokoivula/search-engine.
SearchEngine core JS files are bundled with parcel.js and theme styles are minified with clean-css. Both of these require minimal configuration, which is why all that is needed is the "build" script defined in package.json. To build production assets for SearchEngine:
- Install Parcel globally:
npm install -g parcel-bundler - Install clean-css and clean-css-cli globally:
npm install -g clean-css && npm install -g clean-css-cli - Run build script in the SearchEngine module directory:
npm run build
This project is licensed under the Mozilla Public License Version 2.0.