source.inc in Migrate 6.2

Define base for migration sources.

File

includes/source.inc

<?php

/**
 * @file
 * Define base for migration sources.
 */

/**
 * Abstract base class for source handling.
 *
 * Derived classes are expected to define __toString(), returning a string
 * describing the source and significant options. See
 * MigrateSourceSQL for an example.
 */
abstract class MigrateSource implements Iterator {

  /**
   * The current row from the quey
   *
   * @var stdClass
   */
  protected $currentRow;

  /**
   * The primary key of the current row
   *
   * @var array
   */
  protected $currentKey;
  public function getCurrentKey() {
    return $this->currentKey;
  }

  /**
   * The Migration class currently invoking us, during rewind() and next().
   *
   * @var Migration
   */
  protected $activeMigration;

  /**
   * The MigrateMap class for the current migration.
   *
   * @var MigrateMap
   */
  protected $activeMap;

  /**
   * Number of rows intentionally ignored (prepareRow() returned FALSE)
   *
   * @var int
   */
  protected $numIgnored = 0;
  public function getIgnored() {
    return $this->numIgnored;
  }

  /**
   * Number of rows we've at least looked at.
   *
   * @var int
   */
  protected $numProcessed = 0;
  public function getProcessed() {
    return $this->numProcessed;
  }

  /**
   * Reset numIgnored back to 0.
   */
  public function resetStats() {
    $this->numIgnored = 0;
  }

  /**
   * Information on the highwater mark for the current migration, if any.
   *
   * @var array
   */
  protected $highwaterField;

  /**
   * List of source IDs to process.
   *
   * @var array
   */
  protected $idList = array();

  /**
   * Derived classes must implement fields(), returning a list of available
   * source fields.
   *
   * @return array
   *  Keys: machine names of the fields (to be passed to addFieldMapping)
   *  Values: Human-friendly descriptions of the fields.
   */
  public abstract function fields();

  /**
   * Whether this instance should cache the source count.
   *
   * @var boolean
   */
  protected $cacheCounts = FALSE;

  /**
   * Key to use for caching counts.
   *
   * @var string
   */
  protected $cacheKey;

  /**
   * Whether this instance should not attempt to count the source.
   *
   * @var boolean
   */
  protected $skipCount = FALSE;

  /**
   * By default, next() will directly read the map row and add it to the data
   * row. A source plugin implementation may do this itself (in particular, the
   * SQL source can incorporate the map table into the query) - if so, it should
   * set this TRUE so we don't duplicate the effort.
   *
   * @var bool
   */
  protected $mapRowAdded = FALSE;

  /**
   * Return a count of available source records, from the cache if appropriate.
   * Returns -1 if the source is not countable.
   *
   * @param boolean $refresh
   */
  public function count($refresh = FALSE) {
    if ($this->skipCount) {
      return -1;
    }
    if (!isset($this->cacheKey)) {
      $this->cacheKey = md5((string) $this);
    }

    // If a refresh is requested, or we're not caching counts, ask the derived
    // class to get the count from the source.
    if ($refresh || !$this->cacheCounts) {
      $count = $this
        ->computeCount();
      cache_set($this->cacheKey, $count, 'cache');
    }
    else {

      // Caching is in play, first try to retrieve a cached count.
      $cache_object = cache_get($this->cacheKey, 'cache');
      if (is_object($cache_object)) {

        // Success
        $count = $cache_object->data;
      }
      else {

        // No cached count, ask the derived class to count 'em up, and cache
        // the result
        $count = $this
          ->computeCount();
        cache_set($this->cacheKey, $count, 'cache');
      }
    }
    return $count;
  }

  /**
   * Derived classes must implement computeCount(), to retrieve a fresh count of
   * source records.
   */

  //abstract public function computeCount();

  /**
   * Class constructor.
   *
   * @param array $options
   *  Optional array of options.
   */
  public function __construct($options = array()) {
    if (!empty($options['cache_counts'])) {
      $this->cacheCounts = TRUE;
    }
    if (!empty($options['skip_count'])) {
      $this->skipCount = TRUE;
    }
    if (!empty($options['cache_key'])) {
      $this->cacheKey = $options['cache_key'];
    }
  }

  /**
   * Default implementations of Iterator methods - many derivations will find
   * these adequate and will only need to implement rewind() and next()
   */

  /**
   * Implementation of Iterator::current() - called when entering a loop
   * iteration, returning the current row
   */
  public function current() {
    return $this->currentRow;
  }

  /**
   * Implementation of Iterator::key - called when entering a loop iteration, returning
   * the key of the current row. It must be a scalar - we will serialize
   * to fulfill the requirement, but using getCurrentKey() is preferable.
   */
  public function key() {
    return serialize($this->currentKey);
  }

  /**
   * Implementation of Iterator::valid() - called at the top of the loop, returning
   * TRUE to process the loop and FALSE to terminate it
   */
  public function valid() {
    return !is_null($this->currentRow);
  }

  /**
   * Implementation of Iterator::rewind() - subclasses of MigrateSource should
   * implement performRewind() to do any class-specific setup for iterating
   * source records.
   */
  public function rewind() {
    $this->activeMigration = Migration::currentMigration();
    $this->activeMap = $this->activeMigration
      ->getMap();
    $this->numProcessed = 0;
    $this->numIgnored = 0;
    $this->highwaterField = $this->activeMigration
      ->getHighwaterField();
    if ($this->activeMigration
      ->getOption('idlist')) {
      $this->idList = explode(',', $this->activeMigration
        ->getOption('idlist'));
    }
    else {
      $this->idList = array();
    }
    migrate_instrument_start(get_class($this) . ' performRewind');
    $this
      ->performRewind();
    migrate_instrument_stop(get_class($this) . ' performRewind');
    $this
      ->next();
  }

  /**
   * Implementation of Iterator::next() - subclasses of MigrateSource should
   * implement getNextRow() to retrieve the next valid source rocord to process.
   */
  public function next() {
    $this->currentKey = NULL;
    $this->currentRow = NULL;
    migrate_instrument_start(get_class($this) . ' getNextRow');
    while ($row = $this
      ->getNextRow()) {
      migrate_instrument_stop(get_class($this) . ' getNextRow');

      // Populate the source key for this row
      $this->currentKey = $this->activeMigration
        ->prepareKey($this->activeMap
        ->getSourceKey(), $row);

      // Pick up the existing map row, if any, unless getNextRow() did it.
      if (!$this->mapRowAdded) {
        $map_row = $this->activeMap
          ->getRowBySource($this->currentKey);

        // Add map info to the row, if present
        if ($map_row) {
          foreach ($map_row as $field => $value) {
            $field = 'migrate_map_' . $field;
            $row->{$field} = $value;
          }
        }
      }

      // First, determine if this row should be passed to prepareRow(), or skipped
      // entirely. The rules are:
      // 1. If there's an explicit idlist, that's all we care about (ignore
      //    highwaters and map rows).
      $prepared = FALSE;
      if (!empty($this->idList)) {
        if (in_array(reset($this->currentKey), $this->idList)) {

          // In the list, fall through.
        }
        else {

          // Not in the list, skip it
          $this->currentRow = NULL;
          continue;
        }
      }
      elseif (!isset($row->migrate_map_sourceid1)) {

        // Fall through
      }
      elseif ($row->migrate_map_needs_update == MigrateMap::STATUS_NEEDS_UPDATE) {

        // Fall through
      }
      elseif (empty($this->highwaterField)) {

        // No highwater, skip
        $this->currentRow = NULL;
        continue;
      }
      elseif ($this->activeMigration
        ->getHighwater() === '') {

        // Fall through
      }
      else {

        // Call prepareRow() here, in case the highwaterField needs preparation
        if ($this
          ->prepareRow($row) !== FALSE) {
          if ($row->{$this->highwaterField['name']} > $this->activeMigration
            ->getHighwater()) {
            $this->currentRow = $row;
            break;
          }
          else {

            // Skip
            $this->currentRow = NULL;
            continue;
          }
        }
        else {
          $this->currentRow = NULL;
        }
        $prepared = TRUE;
      }

      // Allow the Migration to prepare this row. prepareRow() can return boolean
      // FALSE to ignore this row.
      if (!$prepared) {
        if ($this
          ->prepareRow($row) !== FALSE) {

          // Finally, we've got a keeper.
          $this->currentRow = $row;
          break;
        }
        else {
          $this->currentRow = NULL;
        }
      }
    }
    migrate_instrument_stop(get_class($this) . ' getNextRow');
    if (!$this->currentRow) {
      $this->currentKey = NULL;
    }
  }

  /**
   * Give the calling migration a shot at manipulating, and possibly rejecting,
   * the source row.
   *
   * @return bool
   *  FALSE if the row is to be skipped.
   */
  protected function prepareRow($row) {
    migrate_instrument_start(get_class($this->activeMigration) . ' prepareRow');
    $return = $this->activeMigration
      ->prepareRow($row);
    migrate_instrument_stop(get_class($this->activeMigration) . ' prepareRow');

    // We're explicitly skipping this row - keep track in the map table
    if ($return === FALSE) {
      $this->activeMigration
        ->getMap()
        ->saveIDMapping($row, array(), MigrateMap::STATUS_IGNORED, $this->activeMigration->rollbackAction);
      $this->numIgnored++;
      $this->currentRow = NULL;
      $this->currentKey = NULL;
    }
    else {
      $return = TRUE;
    }
    $this->numProcessed++;
    return $return;
  }

}

Classes

Name	Description
MigrateSource	Abstract base class for source handling.