Developer guide

Apache Solr API

Helper module for developers to interact with apachesolr backend in Drupal.

Usage

Implement hook_apachesolr_api_config() in your module.
It is recommended to declare dependency on yaml_parser and pick up configuration from a yaml file as shown in the example below.

/**
 * Implements hook_apachesolr_api_config().
 */
function rmh_cv_search_apachesolr_api_config() {
  $data = array();
  $path = drupal_get_path('module', 'rmh_cv_search');
  if ($path) {
    $filepath = $path . '/apachesolr_api_config.yaml';
    $realpath = drupal_realpath($filepath);
    $yaml = file_get_contents($filepath);
    $data = yaml_parser_parse_string($yaml);
  }
  return $data;
}

A sample yaml file could look like as follows,

global_search_url: 'sitewide-search'
panelized content types: 
  - article 
  - page 
  - resource 
pages: 
  search_detail_page: 
    url: search-results/% 
    menu_item: 
      title: Search results page 
      callback: 'apachesolr_api_search_page_callback' 
      page arguments: 
        - search_detail_page 
        - 1 
      access arguments: 
        - access content 
      apachesolr: 
        fuzzy: FALSE 
        params: 
          fl: "id, label, content, teaser, entity_id, bundle, entity_type, path, url" 
          hl: TRUE 
          hl.fl: "content, label" 
          rows: 10 
          hl.snippets: 1 
          hl.fragsize: 300 
          hl.simple.pre: "<mark>" 
          hl.simple.post: "</mark>" 
          hl.simple.snippets: TRUE 
          qf: 
            - "label^20" 
            - "content^5" 
          map: 
            label: label 
            content: content 
            url: path 
  autocomplete: 
    url: ajax/rmh-search/% 
    menu_item: 
      title: Search Suggestions page 
      callback: 'apachesolr_api_search_page_callback' 
      page arguments: 
        - autocomplete 
        - 2 
      access arguments: 
        - access content 
      apachesolr: 
        fuzzy: FALSE 
        params: 
          fl: "id, label, content, teaser, entity_id, bundle, entity_type, path, url" 
          hl: TRUE 
          hl.fl: "content, label" 
          rows: 10 
          hl.simple.pre: "<mark>" 
          hl.simple.post: "</mark>" 
          hl.simple.snippets: TRUE 
          qf: 
            - "label^20" 
            - "content^5" 
          map: 
            label: label 
            content: content 
            url: path 
  bookmarked_and_liked: 
    apachesolr: fuzzy: FALSE 
    params: 
      fl: "entity_id id" 
      rows: 4 
      fq: 
        - bundle:resource 
      map: 
        nid: entity_id 
    flag: 
      save_for_later: 
        entity: 
          node: 
            bundle: 
              - resource 
        flag_machine_name: save_for_later 
      recommend: 
        entity: 
          node: 
            bundle: 
              - resource 
        flag_machine_name: like 
    custom_facets: 
      like: 
        label: My likes 
        type: like 
        solr_field: is_flag_like 
        category: current_user 
      bookmark: 
        label: My bookmarks 
        type: bookmark 
        solr_field: im_flag_save_for_later 
        category: current_user 

Configuration Parameters

global_search_url

apachesolr_api provides a form (apachesolr_api_form) for sitewide search implementation.
The search page url can be passed through global_search_url.

panelized content types

Since the panelized nodes don't store their values as fields attached to nodes, we have to do some extra processing to get them indexed.
Pass a list of panelized content types where such processing is required.

pages

Depending on the parameters passed, it would create full search pages, autocomplete path or just simply a config that holds apachesolr parameters for custom queries.
url along with menu_item would be used to create a hook_menu entry.
menu_item will be passed to hook_menu item declaration as is.
apachesolr holds the configuration for the query to be executed.
apachesolr.fuzzy if set to true, it will support fuzzy logic on query supplied.
apachesolr.params will be merged with custom query. Refer to apachesolr documentation for supported parameters.
map would be used to preprocess the results so that fieldnames in apachesolr index would be replaced with custom ones.
Example configuration

search_detail_page: url: search-results/% menu_item: title: Search results page callback: 'rmh_search_search_page_callback' page arguments: - search_detail_page - 1 access arguments: - access content apachesolr: fuzzy: FALSE params: fl: "id, label, content, teaser, entity_id, bundle, entity_type, path, url" hl: TRUE hl.fl: "content, label" rows: 10 hl.snippets: 1 hl.fragsize: 300 hl.simple.pre: "<mark>" hl.simple.post: "</mark>" hl.simple.snippets: TRUE qf: - "label^20" - "content^5" map: label: label content: content url: path 

NOTE:

• For a full page display one could make use of rmh_search_search_page_callback callback provided by the module.
  • Supported types are autocomplete and search_detail_page. This has to be supplied as first parameter in the page arguments for menu_item.
  • search_detail_page Templates associated with the ouptputs are apachesolr-api-results.tpl.php and apachesolr-api-results-empty.tpl.php.
  • autocomplete will use drupal_json_output to give back results.
• You could also use the helper function apachesolr_api_fetch to get the results directly.
  Eg.

global $user; $config_type = 'bookmarked_and_liked'; $query = '*'; $params = [ 'sort' => 'ds_field_mi_created_date asc', 'fq' => ['im_flag_recommend:' . $user->uid], ]; $solr_results = apachesolr_api_fetch($config_type, $query, $params); 

flag

Flag module interactions are not stored on fields attached to entities. If we need to query results based on flaged status, we would need to index them.
Flag data will be attached to the entity with custom fields with following prefixes,

• is_flag_ : Single valued field which will hold total flag count.
• im_flag_ : Multivalued field that holds user id of each user who has flagged the content.
  Eg.

flag: save_for_later: entity: node: bundle: - resource flag_machine_name: save_for_later 

custom_facets

Custom fields added to document index won't expose facets like fields do. Use this configuration to create pseudo facet links for such fields.
apachesolr_api_get_current_user_flagged_filters() will provide you with a list of such pseudo-facet links.
Eg.

custom_facets: like: label: My likes type: like solr_field: is_flag_like category: current_user