Club Drupal

<?php

/**

 * @file

 * Contains FeedsJSONPathParser.

 */

/**

 * Parses JSON using JSONPath.

 */

class FeedsJSONPathParser extends FeedsParser {

  /**

   * A regular expression that finds four byte UTF-8 chars.

   *

   * @var string

   */

  protected static $fourByteRegex = '/(?:\xF0[\x90-\xBF][\x80-\xBF]{2}|[\xF1-\xF3][\x80-\xBF]{3}|\xF4[\x80-\x8F][\x80-\xBF]{2})/s';

  /**

   * The source fields to debug.

   *

   * @var array

   */

  protected $debug = array();

  /**

   * Implements FeedsParser::parse().

   */

  public function parse(FeedsSource $source, FeedsFetcherResult $fetcher_result) {

    $mappings = $this->getOwnMappings();

    $source_config = $source->getConfigFor($this);

    // Allow config inheritance.

    if (empty($source_config)) {

      $source_config = $this->config;

    }

    $this->debug = array_keys(array_filter($source_config['debug']['options']));

    $raw = trim($fetcher_result->getRaw());

    $result = new FeedsParserResult();

    // Set link so we can set the result link attribute.

    $fetcher_config = $source->getConfigFor($source->importer->fetcher);

    $result->link = $fetcher_config['source'];

    $array = json_decode($raw, TRUE);

    // Support JSON lines format.

    if (!is_array($array)) {

      $raw = preg_replace('/}\s*{/', '},{', $raw);

      $raw = '[' . $raw . ']';

      $array = json_decode($raw, TRUE);

    }

    if (!is_array($array)) {

      throw new Exception(t('There was an error decoding the JSON document.'));

    }

    require_once feeds_jsonpath_parser_library_path();

    $all_items = $this->jsonPath($array, $source_config['context']);

    unset($array);

    // Batch.

    $state = $source->state(FEEDS_PARSE);

    if (!$state->total) {

      $state->total = count($all_items);

    }

    $start = (int) $state->pointer;

    $state->pointer = $start + $source->importer->getLimit();

    $all_items = array_slice($all_items, $start, $source->importer->getLimit());

    // Set progress state.

    $state->progress($state->total, $state->pointer);

    // Debug output.

    $this->debug($all_items, 'context');

    foreach ($all_items as $item) {

      // Invoke a hook to check whether the item should be skipped.

      if ($this->invokeHook($item, $source) === TRUE) {

        continue;

      }

      $parsed_item = $variables = array();

      foreach ($source_config['sources'] as $source_key => $query) {

        // Variable substitution.

        $query = strtr($query, $variables);

        $parsed = $this->parseSourceElement($item, $query, $source_key);

        $variables['{' . $mappings[$source_key] . '}'] = is_array($parsed) ? reset($parsed) : $parsed;

        // Avoid null values.

        if (isset($parsed)) {

          $parsed_item[$source_key] = $parsed;

        }

      }

      if (!empty($parsed_item)) {

        $result->items[] = $parsed_item;

      }

    }

    return $result;

  }

  /**

   * Utilizes the jsonPath function from jsonpath-0.8.1.php.

   *

   * jsonPath returns false if the expression returns zero results and that will

   * mess up our for loops, so return an empty array instead.

   *

   * @param array $array

   *   The input array to parse.

   * @param string $expression

   *   The JSONPath expression.

   *

   * @return array

   *   Returns an array that is the output of jsonPath.

   *

   * @todo

   *   Firgure out error handling.

   */

  protected function jsonPath($array, $expression) {

    $result = jsonPath($array, $expression);

    return ($result === FALSE) ? array() : $result;

  }

  /**

   * Parses one item from the context array.

   *

   * @param array $item

   *   An array containing one item from the context.

   * @param string $query

   *   A JSONPath query.

   * @param string $source

   *   The source element that corresponds to the query.

   *

   * @return array

   *   An array containing the results of the query.

   */

  protected function parseSourceElement($item, $query, $source) {

    if (empty($query)) {

      return;

    }

    $results = $this->jsonPath($item, $query);

    $this->debug($results, $source);

    $count = count($results);

    if ($count === 0) {

      return;

    }

    foreach ($results as $delta => $value) {

      if (is_string($value) && $value !== '') {

        $results[$delta] = !empty($this->config['convert_four_byte']) ? $this->convertFourBytes($value) : $this->stripFourBytes($value);

      }

    }

    if ($count === 1) {

      return reset($results);

    }

    return $results;

  }

  /**

   * Source form.

   */

  public function sourceForm($source_config) {

    $form = array();

    if (empty($source_config)) {

      $source_config = $this->config;

    }

    if (isset($source_config['allow_override']) &&

        !$source_config['allow_override'] &&

        empty($source_config['config'])) {

      return;

    }

    // Add extensions that might get importerd.

    $fetcher = feeds_importer($this->id)->fetcher;

    if (isset($fetcher->config['allowed_extensions'])) {

      if (strpos($fetcher->config['allowed_extensions'], 'json') === FALSE) {

        $fetcher->config['allowed_extensions'] .= ' json';

      }

    }

    $mappings_ = feeds_importer($this->id)->processor->config['mappings'];

    $uniques = $mappings = array();

    foreach ($mappings_ as $mapping) {

      if (strpos($mapping['source'], 'jsonpath_parser:') === 0) {

        $mappings[$mapping['source']] = $mapping['target'];

        if (!empty($mapping['unique'])) {

          $uniques[] = $mapping['target'];

        }

      }

    }

    $form['jsonpath'] = array(

      '#type' => 'fieldset',

      '#title' => t('JSONPath Parser Settings'),

      '#collapsible' => TRUE,

      '#collapsed' => TRUE,

      '#tree' => TRUE,

    );

    if (empty($mappings)) {

      // Detect if Feeds menu structure has changed. This will take a while to

      // be released, but since I run dev it needs to work.

      $feeds_menu = feeds_ui_menu();

      if (isset($feeds_menu['admin/structure/feeds/list'])) {

        $feeds_base = 'admin/structure/feeds/edit/';

      }

      else {

        $feeds_base = 'admin/structure/feeds/';

      }

      $form['jsonpath']['error_message']['#markup'] = '<div class="help">' . t('No JSONPath mappings are defined. Define mappings !link.', array('!link' => l(t('here'), $feeds_base . $this->id . '/mapping'))) . '</div><br />';

      return $form;

    }

    $form['jsonpath']['context'] = array(

      '#type' => 'textfield',

      '#title' => t('Context'),

      '#required' => TRUE,

      '#description' => t('This is the base query, all other queries will execute in this context.'),

      '#default_value' => isset($source_config['context']) ? $source_config['context'] : '',

      '#maxlength' => 1024,

      '#size' => 80,

    );

    $form['jsonpath']['sources'] = array(

      '#type' => 'fieldset',

    );

    if (!empty($uniques)) {

      $items = array(

        format_plural(count($uniques),

          t('Field <strong>!column</strong> is mandatory and considered unique: only one item per !column value will be created.',

            array('!column' => implode(', ', $uniques))),

          t('Fields <strong>!columns</strong> are mandatory and values in these columns are considered unique: only one entry per value in one of these columns will be created.',

            array('!columns' => implode(', ', $uniques)))),

      );

      $form['jsonpath']['sources']['help']['#markup'] = '<div class="help">' . theme('item_list', array('items' => $items)) . '</div>';

    }

    $variables = array();

    foreach ($mappings as $source => $target) {

      $form['jsonpath']['sources'][$source] = array(

        '#type' => 'textfield',

        '#title' => $target,

        '#description' => t('The JSONPath expression to execute.'),

        '#default_value' => isset($source_config['sources'][$source]) ? $source_config['sources'][$source] : '',

        '#maxlength' => 1024,

        '#size' => 80,

      );

      if (!empty($variables)) {

        $variable_text = format_plural(count($variables),

          t('The variable %v is available for replacement.', array('%v' => implode(', ', $variables))),

          t('The variables %v are available for replacement.', array('%v' => implode(', ', $variables)))

        );

        $form['jsonpath']['sources'][$source]['#description'] .= '<br />' . $variable_text;

      }

      $variables[] = '{' . $target . '}';

    }

    $form['jsonpath']['debug'] = array(

      '#type' => 'fieldset',

      '#title' => t('Debug'),

      '#collapsible' => TRUE,

      '#collapsed' => TRUE,

    );

    $form['jsonpath']['debug']['options'] = array(

      '#type' => 'checkboxes',

      '#title' => t('Debug query'),

      '#options' => array_merge(array('context' => 'context'), $mappings),

      '#default_value' => isset($source_config['debug']['options']) ? $source_config['debug']['options'] : array(),

    );

    return $form;

  }

  /**

   * Override parent::configForm().

   */

  public function configForm(&$form_state) {

    $config = $this->getConfig();

    $config['config'] = TRUE;

    $form = $this->sourceForm($config);

    $form['jsonpath']['context']['#required'] = FALSE;

    $form['jsonpath']['#collapsed'] = FALSE;

    $form['jsonpath']['allow_override'] = array(

      '#type' => 'checkbox',

      '#title' => t('Allow source configuration override'),

      '#description' => t('This setting allows feed nodes to specify their own JSONPath values for the context and sources.'),

      '#default_value' => $config['allow_override'],

    );

    $form['jsonpath']['convert_four_byte'] = array(

      '#type' => 'checkbox',

      '#title' => t('Convert four byte characters'),

      '#description' => t('Coverts four byte UTF-8 characters to their HTML entity. By default, four byte characters will be stripped.'),

      '#default_value' => !empty($config['convert_four_byte']),

    );

    return $form;

  }

  /**

   * Override parent::getMappingSources().

   */

  public function getMappingSources() {

    $mappings = $this->filterMappings(feeds_importer($this->id)->processor->config['mappings']);

    $next = 0;

    if (!empty($mappings)) {

      $keys = array_keys($mappings);

      $nums = array();

      foreach ($keys as $key) {

        list(, $num) = explode(':', $key);

        $nums[] = $num;

      }

      $max = max($nums);

      $next = ++$max;

    }

    return array(

      'jsonpath_parser:' . $next => array(

        'name' => t('JSONPath Expression'),

        'description' => t('Allows you to configure a JSONPath expression that will populate this field.'),

      ),

    ) + parent::getMappingSources();

  }

  public function sourceDefaults() {

    return array();

  }

  /**

   * Define defaults.

   */

  public function configDefaults() {

    return array(

      'context' => '',

      'sources' => array(),

      'debug' => array(),

      'allow_override' => FALSE,

      'convert_four_byte' => FALSE,

    );

  }

  /**

   * Override parent::sourceFormValidate().

   *

   * If the values of this source are the same as the base config we set them to

   * blank to that the values will be inherited from the importer defaults.

   *

   * @param array $values

   *   The values from the form to validate, passed by reference.

   */

  public function sourceFormValidate(&$values) {

    $config = $this->getConfig();

    $values = $values['jsonpath'];

    $allow_override = $config['allow_override'];

    unset($config['allow_override']);

    unset($config['convert_four_byte']);

    ksort($values);

    ksort($config);

    if ($values === $config || !$allow_override) {

      $values = array();

      return;

    }

    $this->configFormValidate($values);

  }

  /**

   * Override parent::sourceFormValidate().

   */

  public function configFormValidate(&$values) {

    if (isset($values['jsonpath'])) {

      $values = $values['jsonpath'];

    }

    $values['context'] = isset($values['context']) ? trim($values['context']) : '';

    if (!empty($values['sources'])) {

      foreach ($values['sources'] as &$source) {

        $source = trim($source);

      }

    }

  }

  /**

   * Gets the mappings that belong to this parser.

   *

   * @return array

   *   An array of mappings keyed source => target.

   */

  protected function getOwnMappings() {

    $importer_config = feeds_importer($this->id)->getConfig();

    return $this->filterMappings($importer_config['processor']['config']['mappings']);

  }

  /**

   * Filters mappings, returning the ones that belong to us.

   *

   * @param array $mappings

   *   A mapping array from a processor.

   *

   * @return array

   *   An array of mappings keyed source => target.

   */

  protected function filterMappings($mappings) {

    $our_mappings = array();

    foreach ($mappings as $mapping) {

      if (strpos($mapping['source'], 'jsonpath_parser:') === 0) {

        $our_mappings[$mapping['source']] = $mapping['target'];

      }

    }

    return $our_mappings;

  }

  protected function debug($item, $source) {

    if (in_array($source, $this->debug)) {

      $o = '<ul>';

      foreach ($item as $i) {

        $o .= '<li>' . check_plain(var_export($i, TRUE)) . '</li>';

      }

      $o .= '</ul>';

      drupal_set_message($source . ':' . $o);

    }

  }

  /**

   * Calls our filter hook.

   *

   * @param array &$item

   *   The item to alter.

   * @param FeedsSource $source

   *   The feed source.

   * @return true|null

   *   Returns true if the item should be skipped.

   */

  protected function invokeHook(array &$item, FeedsSource $source) {

    foreach (module_implements('feeds_jsonpath_parser_filter') as $module) {

      $function = $module . '_feeds_jsonpath_parser_filter';

      if ($function($item, $source) === TRUE) {

        return TRUE;

      }

    }

  }

  /**

   * Strips four byte characters from a string.

   *

   * @param string $string

   *   The input string.

   *

   * @return string

   *   The string with four byte characters removed.

   */

  public static function stripFourBytes($string) {

    return preg_replace(self::$fourByteRegex, '', $string);

  }

  /**

   * Replaces four byte characters with their HTML unicode codepoint.

   *

   * @param string $string

   *   The input string.

   *

   * @return string

   *   The string with four byte characters converted.

   */

  public static function convertFourBytes($string) {

    return preg_replace_callback(self::$fourByteRegex, array('FeedsJSONPathParser', 'doFourByteReplace'), $string);

  }

  /**

   * Callback for FeedsJSONPathParser::convertFourBytes().

   *

   * @param array $matches

   *   The regular expression matches.

   *

   * @return string

   *   A four byte unicode character converted to its HTML representation.

   */

  public static function doFourByteReplace(array $matches) {

    $char = $matches[0];

    // Calculate the codepoint of the character.

    $codepoint = ord($char[0]) - 0xF0 << 18;

    $codepoint += ord($char[1]) - 0x80 << 12;

    $codepoint += ord($char[2]) - 0x80 << 6;

    $codepoint += ord($char[3]) - 0x80;

    return '&#' . $codepoint . ';';

  }

}