<?php
// $Id$

/**
 * @file
 * Bulk import script for Calais.
 *
 * Use this when you have enough content on your site that it would be
 * completely silly to sit around and wait for the batch API import process
 * to complete.
 *
 * To run, move this file into your Drupal root directory and either run it
 * in the browser, or invoke it in a crob job.
 */

/**
 * How many nodes to process per batch run.
 */
define('CALAIS_BATCH_LIMIT', 10);

// Bootstrap Drupal.
require_once './includes/bootstrap.inc';
drupal_bootstrap(DRUPAL_BOOTSTRAP_FULL);

// Run the batch import process.
calais_bulk_import();

/**
 * Bulk-tag nodes in batches of CALAIS_BATCH_LIMIT.
 *
 * We have to define this here because the calais_batch_process() function
 * assumes batch API.
 */
function calais_bulk_import($throttle = TRUE) {
  // If we're already done, nothing to do here.
  if (variable_get('calais_bulk_import_nodes_completed', FALSE)) {
    return;
  }

  // Get a list of only the node types that have been set to be processed on
  // each update.
  $types = array_keys(node_get_types());
  $valid_types = array();
  foreach ($types as $type) {
    $process_type = variable_get("calais_node_{$type}_process", CALAIS_PROCESS_NO);
    if ($process_type == CALAIS_PROCESS_AUTO) {
      $valid_types[] = $type;
    }
  }
  if (empty($valid_types)) {
    echo 'No content types are Calais-enabled.';
    // Nothing to do here.
    return;
  }

  // Process nodes in ascending order, starting with the last one we processed.
  $last_processed = variable_get('calais_bulk_import_last_processed', 0);
  if (!$last_processed) {
    // Determine the max node ID.
    $last_processed = db_result(db_query("SELECT n.nid FROM {node} n LEFT JOIN {calais_term_node} c ON n.nid = c.nid WHERE n.type IN ('%s') AND c.nid IS NULL ORDER BY n.nid DESC", implode("','", $valid_types)));
  }
  $query = db_query_range("SELECT DISTINCT(n.nid) FROM {node} n LEFT JOIN {calais_term_node} c ON n.nid = c.nid WHERE n.type IN ('%s') AND c.nid IS NULL AND n.nid < %d ORDER BY n.nid DESC", implode("','", $valid_types), $last_processed, 0, CALAIS_BATCH_LIMIT);
  while ($nid = db_result($query)) {
    $node = node_load($nid, NULL, TRUE);
    // Add a flag to indicate we're in the middle of a process so that other
    // modules may react differently if they need to.
    $node->calais_bulk_process = TRUE;
    $threshold = variable_get("calais_threshold_{$node->type}", 0.01);
    calais_process_node($node, $process_type, $threshold, 'update');

    // It's nice to give Calais's poor servers a break in between processing.
    if ($throttle) {
      usleep(100000);
    }
  }

  // Update the last node imported and provide status report.
  variable_set('calais_bulk_import_last_processed', $node->nid);
  $remaining = db_result(db_query("SELECT COUNT(DISTINCT(n.nid)) FROM {node} n LEFT JOIN {calais_term_node} c ON n.nid = c.nid WHERE n.type IN ('%s') AND n.nid < %d AND c.nid IS NULL" , implode("','", $valid_types), $node->nid));
  if ($remaining == 0) {
    // We're done! Hooray!!
    variable_set('calais_bulk_import_nodes_completed', TRUE);
  }
  print t('Processed through record @record. @remaining remaining.', array('@record' => $node->nid, '@remaining' => $remaining));

  // Clear the caches. Ouch. :(
  cache_clear_all();
}