Updates publication records that currently exist in the Chado pub table with the most recent data in the remote database.

Parameters

$do_contact: Set to TRUE if authors should automatically have a contact record added to Chado. Contacts are added using the name provided by the remote database.

$dbxref: The unique database ID for the record to update. This value must be of the format DB_NAME:ACCESSION where DB_NAME is the name of the database (e.g. PMID or AGL) and the ACCESSION is the unique identifier for the record in the database.

$db: The name of the remote database to update. If this value is provided and no dbxref then all of the publications currently in the Chado database for this remote database will be updated.

Related topics

File

tripal_pub/includes/tripal_pub.pub_importers.inc, line 1823

Management of importers

Code

function tripal_reimport_publications($do_contact = FALSE, $dbxref = NULL, $db = NULL) {

  print "\nNOTE: Loading of publications is performed using a database transaction. \n" .
    "If the load fails or is terminated prematurely then the entire set of \n" .
    "insertions/updates is rolled back and will not be found in the database\n\n";
  $transaction = db_transaction();
  try {

    // get a list of all publications by their Dbxrefs that have supported databases
    $sql = "
      SELECT DB.name as db_name, DBX.accession
      FROM pub P
        INNER JOIN pub_dbxref PDBX ON P.pub_id = PDBX.pub_id
        INNER JOIN dbxref DBX      ON DBX.dbxref_id = PDBX.dbxref_id
        INNER JOIN db DB           ON DB.db_id = DBX.db_id
    ";
    $args = array();
    if ($dbxref and preg_match('/^(.*?):(.*?)$/', $dbxref, $matches)) {
      $dbname = $matches[1];
      $accession = $matches[2];
      $sql .= "WHERE DBX.accession = :accession and DB.name = :dbname ";
      $args[':accession'] = $accession;
      $args[':dbname'] = $dbname;
    }
    elseif ($db) {
      $sql .= " WHERE DB.name = :dbname ";
      $args[':dbname'] = $db;
    }
    $sql .= "ORDER BY DB.name, P.pub_id";
    $results = chado_query($sql, $args);

    $num_to_retrieve = 100;
    $i = 0; // count the number of IDs. When we hit $num_to_retrieve we'll do the query
    $curr_db = ''; // keeps track of the current current database
    $ids = array(); // the list of IDs for the database
    $search = array(); // the search array passed to the search function

    // iterate through the pub IDs
    while ($pub = $results->fetchObject()) {
      $accession = $pub->accession;
      $remote_db = $pub->db_name;

      // here we need to only update publications for databases we support
      $supported_dbs = variable_get('tripal_pub_supported_dbs', array());
      if (!in_array($remote_db, $supported_dbs)) {
        continue;
      }
      $search = array(
        'num_criteria' => 1,
        'remote_db' => $remote_db,
        'criteria' => array(
          '1' => array(
            'search_terms' => "$remote_db:$accession",
            'scope' => 'id',
            'operation' => '',
            'is_phrase' => 0,
          ),
        ),
      );
      $pubs = tripal_get_remote_pubs($remote_db, $search, 1, 0);
      tripal_pub_add_publications($pubs, $do_contact, TRUE);

      $i++;
    }

    // sync the newly added publications with Drupal
    print "Syncing publications with Drupal...\n";
    chado_node_sync_records('pub');

    // if the caller wants to create contacts then we should sync them
    if ($do_contact) {
      print "Syncing contacts with Drupal...\n";
      chado_node_sync_records('contact');
    }
  }
  catch (Exception $e) {
    $transaction->rollback();
    print "\n"; // make sure we start errors on new line
    watchdog_exception('T_pub_import', $e);
    print "FAILED: Rolling back database changes...\n";
    return;
  }
  print "Done.\n";
}