Commit a7e6b9de authored by git's avatar git Committed by Markus Kalkbrenner
Browse files

Issue #3048751 by OanaIlea, gnindl, mkalkbrenner: Ukrainian lemmatization

parent 41f30be6

config/optional/search_api_solr.solr_field_type.text_uk_6_0_0.yml

+4 −6
Changes for config/optional/search_api_solr.solr_field_type.text_uk_6_0_0.yml: 4 added lines, 6 removed lines.
Original line number Diff line number Diff line
@@ -39,9 +39,8 @@ field_type: 
          ignoreCase: true

          words: stopwords_uk.txt

        -

          class: solr.SnowballPorterFilterFactory

          language: Russian

          protected: protwords_uk.txt

          class: solr.MorphologikFilterFactory

          dictionary: ukrainian.dict

        -

          class: solr.RemoveDuplicatesTokenFilterFactory

    -
@@ -71,9 +70,8 @@ field_type: 
          ignoreCase: true

          words: stopwords_uk.txt

        -

          class: solr.SnowballPorterFilterFactory

          language: Russian

          protected: protwords_uk.txt

          class: solr.MorphologikFilterFactory

          dictionary: ukrainian.dict

        -

          class: solr.RemoveDuplicatesTokenFilterFactory

unstemmed_field_type:

config/optional/search_api_solr.solr_field_type.text_uk_7_0_0.yml

+4 −6
Changes for config/optional/search_api_solr.solr_field_type.text_uk_7_0_0.yml: 4 added lines, 6 removed lines.
Original line number Diff line number Diff line
@@ -39,9 +39,8 @@ field_type: 
          ignoreCase: true

          words: stopwords_uk.txt

        -

          class: solr.SnowballPorterFilterFactory

          language: Russian

          protected: protwords_uk.txt

          class: solr.MorphologikFilterFactory

          dictionary: ukrainian.dict

        -

          class: solr.RemoveDuplicatesTokenFilterFactory

    -
@@ -71,9 +70,8 @@ field_type: 
          ignoreCase: true

          words: stopwords_uk.txt

        -

          class: solr.SnowballPorterFilterFactory

          language: Russian

          protected: protwords_uk.txt

          class: solr.MorphologikFilterFactory

          dictionary: ukrainian.dict

        -

          class: solr.RemoveDuplicatesTokenFilterFactory

unstemmed_field_type:

config/schema/solr_field_type.schema.yml

+1 −1
Changes for config/schema/solr_field_type.schema.yml: 1 added line, 1 removed line.
Original line number Diff line number Diff line
@@ -246,7 +246,7 @@ search_api_solr.solr_field_type_filters: 
        nullable: true

      dictionary:

        type: string

        label: 'OpenNLPLemmatizerFilter:: HunspellStemFilter:: DictionaryCompoundWordTokenFilter::dictionary'

        label: 'OpenNLPLemmatizerFilter:: HunspellStemFilter:: MorphologikFilterFactory:: DictionaryCompoundWordTokenFilter::dictionary'

        nullable: true

      lemmatizerModel:

        type: string

search_api_solr.install

+29 −0
Changes for search_api_solr.install: 29 added lines, 0 removed lines.
Original line number Diff line number Diff line
@@ -1199,3 +1199,32 @@ function search_api_solr_update_8315() { 
    }

  }

}



/**

 * Use Ukrainian lemmatization instead of Russian SnowballPorterFilterFactory.

 */

function search_api_solr_update_8316() {

  foreach (search_api_solr_update_helper_get_field_type_configs() as $field_type_name => $field_type_config) {

    if ('uk' === $field_type_config['field_type_language_code']) {

      if (!empty($field_type_config['field_type'])) {

        foreach ($field_type_config['field_type'] as &$components) {

          if (is_array($components)) {

            foreach ($components as &$analyzers) {

              if (is_array($analyzers)) {

                foreach ($analyzers['filters'] as $key => &$filter) {

                  if ('solr.SnowballPorterFilterFactory' === $filter['class']) {

                    $filter['class'] = 'solr.MorphologikFilterFactory';

                    unset($filter['language']);

                    unset($filter['protected']);

                    $filter['dictionary'] = 'ukrainian.dict';

                  }

                }

              }

            }

          }

        }

        search_api_solr_update_helper_save_field_type_config($field_type_name, $field_type_config);

      }

    }

  }

}