These are chat archives for collectiveaccess/support

Jul 2017
Jul 19 2017 09:37

I have a question regarding tokenizer regex used for indexing and searching in providence. As mentioned in the search.conf file both regex should be same. However, in ' providence/app/lib/core/Plugins/SearchEngine/SqlSearch.php' file there is a difference in initiation of these regex.

    if (!($this->ops_indexing_tokenizer_regex = trim($this->opo_search_config->get('indexing_tokenizer_regex')))) {
        $this-> ops_indexing_tokenizer_regex = "^\pL\pN\pNd/_#\@\&\.";
    if (!($this->ops_search_tokenizer_regex = trim($this->opo_search_config->get('search_tokenizer_regex')))) {
        $this->ops_search_tokenizer_regex = "^\pL\pN\pNd/_#\@\&";

Question 1: Either search tokenizer regex should have dot(.) as well or index tokenizer should exclude it?
Question 2: Shouldn't the initialization values here be the same as specified in the search.conf file?

Actually, we are facing a problem in search results which apparently is linked with the exclusion of dot(.) in search tokenizer. When we make a search(basic or quick), for example CRKC.0027, no results are returned. The reason being the ops_search_tokenizer_regex, which removes dot(.) from the search string, for example 'CRKC.0027' to 'CRKC0027'. Which off-course is not available. Including dot(.) in ops_search_tokenizer_regex however solves the problem. Therefore i am wondering if dot(.) should be added to the search tokenizer regex in search.conf and the above mentioned initialization value of the ops_search_tokenizer_regex?

Jul 19 2017 13:56
@juliaweist We're running 1.5 in production, and 1.7 in a test environment. I can't find the entity class setting anywhere.
Jul 19 2017 19:36
You might need to add it to the list item UI. Manage > Administration > User interfaces > List items
Jul 19 2017 20:11
Found it! Looks like the preferred labels are appearing the way they should. Thanks!