When the CLI Command runs into an error, you need to manually delete the tmp file.

Some white spaces missing in search results after strip_tags.

Hi

We configured lucene search according to the documentation. The crawler has finished successfully and the search template was added with following code:

<?= $this->actions()->render(new \Symfony\Component\HttpKernel\Controller\ControllerReference('LuceneSearchBundle:List:getResult')); ?>

But when we add /search?p=test nothing is shown in the list. We only get empty divs generated by the lucene search layout.

Did we forgot something?

Our config is:

lucene_search:
    enabled: true
    sitemap:
        render: false
    fuzzy_search_results: false
    search_suggestion: true
    seeds:
        - 'http://domain-xy.com'
    filter:
        valid_links:
            - '@^http://domain-xy.com.*@i'
    view:
        max_per_page: 10
    crawler:
        content_max_size: 4
    locale:
        ignore_language: true

This has been dramatically changed in pimcore build 3987 (see pimcore/pimcore#937).

currently if meta tag has been set, lucene search will throw some heavy errors.

Update to latest vdb/php-spider.
Because this is still the best crawl library out there, we may gonna fork and tag this repo.

replace tagfields with something like this: http://examples.sencha.com/extjs/6.0.1/examples/kitchensink/#form-fieldreplicator

if country restriction has been activated, no pdf will be shown up - it's not possible to detect.

Task I

Add assigned_country as asset property, to get associated country

Task II

Allow user to set asset free from country restriction: Update countryQuery and exclude assets from country restriction queries.

This is the message:
Fatal error: Class 'LuceneSearch\Model\Configuration' not found in /var/www/plugins/LuceneSearch/controllers/FrontendController.php on line 48

Andreas

Sorry - installed it by composer and thought that plugin is installed by composer. But now I see that it is necessary to install it also by "activate extension" in pimcore backend.

• respect categories meta tag, if available.
• allow multiple categories
• update readme.md and provide frontend category selection sample

Configuration

lucene_search:
    categories:
        - {id: 1, label: 'Category 1'}
        - {id: 2, label: 'Category 2'}

View

{% if lucene_search_crawler_active() %}
      <meta name="lucene-search:categories">1,2</head>
{% endif %}

I installed for Pimcore 5 using composer . Activated using extension manager
but can't get the crawler to start.

Keep getting error :
no valid task iterators defined!

I start the crawler with :
php bin/console lucenesearch:crawl -v

Any idea what could be going wrong ?

Thanks!

Is there a specific reason why you read the complete log file in order to add a line to it? This seems to be much overhead once the log file grows bigger - why not just something like:

$current = date('d.m.Y H:i') . '|' . $this->getRealLevel($level) . '|' . $message . "\n";
file_put_contents($file, $current, FILE_APPEND);

• maxSuggestions
• perPage

Allow some custom meta data which should be indexed only in crawler mode.

if( \LuceneSearch\Tool\Request::isLuceneSearchCrawler() )
{
       $this->view->headMeta()->setName( 'lucene-search-meta', 'awesome special data' );
}

Multiple Start Urls for crawling result in having only indexed the latest one. For example:

b2b.domain.com
www.domain.com
test.domain.com

LuceneSearch crawls these domains, but only indexes the last one.

If there is no state.cnf, LuceneSearch fails with following error:

invalid state config slot "finished"

This is due to a missing state.cnf. But there is currently no state.cnf, because LuceneSearch support has just been deployed.

It should be possible to display the meta description instead of "random" content in result list.

In some cases (especially large pages), the oom killer triggers because of heavy memory usage.
Use file persistence to avoid this.

Also use renderTemplate for listAction.

I followed the guide to implement the search page with ajax autoComplete
when ,I search, I get the following error :

Symfony\Component\Debug\Exception\ContextErrorException
in vendor\dachcom-digital\lucene-search\src\LuceneSearchBundle\Controller\AutoCompleteController.php (line 26)
foreach ($terms as $term) { $t = $term->text;

Any help ?

Find a good way to localize pdf indexing.

Pimcore has its own Sitemap Builder now: pimcore/pimcore#1901

• Add "enable/disable PDF indexing" to backend configuration
• Add "PDF max file size for indexing" to backend configuration

use a more generic config style:

instead of

restriction:ignore: true
restriction:class: ~
restriction:method: ~

use

restriction:
    ignore: true
    class: ~
    method: ~

I need categories from the Pimcore objects. Unfortunately I don't have access to the class category.

AppBundle\LuceneSearch\Services\Categories

    public function getCategories() : array
    {
        $categoryArray = [];

        array_push($categoryArray, ['id' => 1, 'label' => 'Blog']);
        array_push($categoryArray, ['id' => 2, 'label' => 'Support']);
        array_push($categoryArray, ['id' => 3, 'label' => 'Project']);

        $catList = new DataObject\....\Listing(); // -> ERROR (Fatal error: Uncaught Symfony\Component\Debug\Exception\FatalThrowableError: Call to a member function get() on null in /var/www/html/pimcore/lib/Pimcore/Db.php on line 51) 
        foreach ($catList as $cat) {
            array_push($categoryArray, ['id' => $cat->getId(), 'label' => $cat->getName()]);
        }

        return $categoryArray;
    }

Allow to set some exclude tags. For example:

main-content: Keep Data in between!
main-content-exclude: Remove Data in between!

<!-- main-content-->
<span>Good Content! Vivamus magna justo, lacinia eget consectetur sed, convallis at tellus.</span>
<!-- main-content-exclude -->
<span class="boring-text">Boring! This will be excluded in search!</span>
<!-- /main-content-exclude -->
<span>Good Content! Vivamus magna justo, lacinia eget consectetur sed, convallis at tellus</span>
<!-- /main-content-->

Hello,

I was asking me if the plugin can handle subsite and langage separation (for search, autocomplete and suggestion), so it send only result for the langage or subsite ?
Is there any way to configure it like this ?

If not, maybe it could be improved to handle multi-site and multi-langage case.

Thanks again for this wonderfull plugin very complete (i really liked the auto sugestion : )) like Google)

Add a new feature to show all indexed documents and closer inspection in admin backend, this allows for better debugging and understandability what the index actually saved.

Tried already and had some conflicts with composer since pimcore requires
"zendframework/zend-servicemanager": "^3.2" which relies on the "zendframework/zend-stdlib": "^3.1" package.

Since the ZendSearch packages requires "zendframework/zend-stdlib": "2.*" we're going to run into some conflicts.

Hi, first of all - sorry for re-opening but in my case it needs to be :( I'm running pimcore 5, everything works well and I installed lucene using the composer, activated at the extension manager and so on. I placed the lucene config from above example at src/Appbundle/Resources/config/pimcore/config.yml as suggested but if i start a manual crawl on the console I still having the Error "no valid task iterators defined!"

any ideas ? thanks in advance ! kind regards

• parser: check asset restriction with eventdispatcher
• frontend: check query restriction with eventdispatcher and remove class/method config
• config: if restriction.enabled = false:
• • do not add any restriction terms to lucene
• • skip event dispatching

Add some log infos about the last crawl, including:

• all persisted links
• error logs

the regex @^http://www\\.pimcore\\.org*@i is wrong, should be @^http://www\.pimcore\.org*@i

The instructions state that I should setup a view script containing:
$this->action('find', 'frontend', 'LuceneSearch', array('viewScript' => 'frontend/find.php'));
This yields the following error:

script 'frontend/find.php' not found in path (/Users/patrick/Sites/externt/lucenetest/plugins/LuceneSearch/views/scripts/:/Users/patrick/Sites/externt/lucenetest/website/views/layouts/:/Users/patrick/Sites/externt/lucenetest/website/views/scripts/:./views/layouts/:./views/scripts/)

Just copying the contents of plugins/LuceneSearch/views/scripts/search/find.php won't do since it's useless without it's controller ..

What am I missing?

Sometimes, its very hard to debug Lucene Search. Therefore it would be nice to have more debug informations like:

• I am currently crawling this site: $url
• I am denying this site: $url cause of this regex: $url
• I added this site $site to the index with this $language/$country (or whatever we can print here)

In order to install the Plugin, it must be enabled first. However - once enabled, the class DependencyInjection/LuceneSearchExtension.php expects the config file var/bundles/LuceneSearchBundle/config.yml to alreay exist. This yields a PHP warning, which in turn throws an exception in dev-Environment or when debug mode is enabled.

Steps to reproduce: (PIMCORE_ENVIRONMENT is "dev")

vagrant@develop:/workspace/test$ bin/console pimcore:bundle:enable LuceneSearchBundle
[OK] Bundle "LuceneSearchBundle\LuceneSearchBundle" was successfully enabled

vagrant@develop:/workspace/test$ bin/console pimcore:bundle:install LuceneSearchBundle
PHP Fatal error: Uncaught Symfony\Component\Debug\Exception\ContextErrorException: Warning: file_get_contents(/workspace/test/var/bundles/LuceneSearchBundle/config.yml): failed to open stream: No such file or directory in src/LuceneSearchBundle/DependencyInjection/LuceneSearchExtension.php:26

Proposed fix: Just suppressing the warning using the @ in front of file_get_contents:
$bundleConfig =Yaml::parse(@file_get_contents(BundleConfiguration::SYSTEM_CONFIG_FILE_PATH));