<?php
/*
  Copyright (C) 2008 by Phase2 Technology.
  Author(s): Frank Febbraro, Irakli Nadareishvili

  This program is free software; you can redistribute it and/or modify
  it under the terms of the GNU General Public License.
  This program is distributed in the hope that it will be useful,
  but WITHOUT ANY WARRANTY. See the LICENSE.txt file for more details.

  $Id: calais_api.module,v 1.3.2.10.2.9 2009-01-22 20:23:01 febbraro Exp $
*/
/**
 * @file
 */

module_load_include('inc', 'calais_api', 'calais.test');
module_load_include('inc', 'calais_api', 'includes/Calais');
module_load_include('inc', 'calais_api', 'includes/CalaisMetadata');

/**
 * Implementation of hook_perm().
 */
function calais_api_perm() {
  return array('administer calais api');
}

/**
 * Implementation of hook_menu().
 */
function calais_api_menu() {
  $items = array();
  
  $items['admin/settings/calais'] = array(
    'title' => 'Calais Configuration',
    'description' => 'Configurations for Calais',
    'page callback' => 'drupal_get_form',
    'page arguments' => array('calais_api_admin_settings'),
    'access arguments' => array('administer calais api'),
  );

  $items['admin/settings/calais/calais-api'] = array(
    'title' => 'Calais API Settings',
    'description' => 'Configurations for Calais API',
    'type' => MENU_DEFAULT_LOCAL_TASK,
  );

  // Used for testing/debug
  $items['admin/settings/calais/test'] = array(
    'title' => 'Calais Tester Page',
    'description' => 'Calais Test Call',
    'page callback' => 'calais_call_test',
    'access arguments' => array('administer calais api'),
    'type' => MENU_CALLBACK,
  );

  return $items;
}

/**
 * Build the admin settings form.
 */
function calais_api_admin_settings() {
  $form = array();
  $calais_url = array(
    '!calaisurl' => l(t('Calais Website'), 'http://www.opencalais.com/user/register')
    );
  
  $form['calais_api_key'] = array(
    '#type' => 'textfield',
    '#title' => t('Calais API Key'),
    '#default_value' => variable_get('calais_api_key', NULL),
    '#size' => 60,
    '#description' => t('You need to obtain an API Key from the !calaisurl first', $calais_url),
  );
  $form['calais_api_allow_searching'] = array(
    '#type' => 'checkbox',
    '#title' => t('Allow Calais Searching'),
    '#default_value' => variable_get('calais_api_allow_searching', TRUE),
    '#description' => t('Indicates whether future searches can be performed on the extracted metadata by Calais'),
  );
  $form['calais_api_allow_distribution'] = array(
    '#type' => 'checkbox',
    '#title' => t('Allow Calais Distribution'),
    '#default_value' => variable_get('calais_api_allow_distribution', TRUE),
    '#description' => t('Indicates whether the extracted metadata can be distributed by Calais'),
  );
  
  $form = system_settings_form($form);
  return $form;
}

/**
 * Analyze the content via Calais.
 *
 * @param $content The content to ship off to Calais for analysis
 * @param $parameters Array of Calais parameters for overriding defaults.
 * @see http://www.opencalais.com/APIcalls#inputparameters
 *
 * @return The analyzed content
 */
function calais_api_analyze($content, $parameters = array()) {
  $calais = new Calais($parameters);
  return $calais->analyze($content);
}

/**
 * Analyze the content via the Calais XML interface.
 *
 * @param $title The title of the content
 * @param $body The entire body of the content
 * @param $date The date of the content (can be created or updated date) 
 *    This date is used to base calculations of words like "tomorrow" or "yesterday"
 * @param $parameters Array of Calais parameters for overriding defaults.
 * @see http://www.opencalais.com/APIcalls#inputparameters
 *
 * @return The analyzed content
 */
function calais_api_analyze_xml($title, $body, $date, $parameters = array()) {
  $calais = new Calais($parameters);
  return $calais->analyzeXML($title, $body, $date);
}

/**
 * Takes a CamelCase word and adds spaces to make it Camel Case
 *
 * @return an formated string
 */
function calais_api_make_readable($camel_case) {
  return preg_replace('/(.*?[a-z]{1})([A-Z]{1}.*?)/', '${1} ${2}', $camel_case);
}

/**
 * Declare the Calais namespace to the RDF API module (in case we use it)
 */
function calais_api_rdf_namespaces() {
  return array(
    'c' => 'http://s.opencalais.com/1/pred/',
    'sys' => 'http://s.opencalais.com/1/type/sys/',
    'lid' => 'http://s.opencalais.com/1/type/lid/',
    'cat' => 'http://s.opencalais.com/1/type/cat/',
    'resolved' => 'http://s.opencalais.com/1/type/er/',
    'eventfact' => 'http://s.opencalais.com/1/type/em/r/',
    'entity' => 'http://s.opencalais.com/1/type/em/e/',
  );
}

/**
 * Get a list of the entities that Calais API defines:
 * http://opencalais.mashery.com/page/calaissemanticmetadata
 *
 * TODO: When Calais updates to have a static list at a URL or via API call, return that instead.
 *
 * @return flat array listing of Calais entities
 */
function calais_api_get_all_entities() {
  return array(
    'Anniversary',
    'City',
    'Company',
    'Continent',
    'Country',
    'Currency',
    'EmailAddress',
    'EntertainmentAwardEvent',
    'Facility',
    'FaxNumber',
    'Holiday',
    'IndustryTerm',
    'MarketIndex',
    'MedicalCondition',
    'MedicalTreatment',
    'Movie',
    'MusicAlbum',
    'MusicGroup',
    'NaturalDisaster',
    'NaturalFeature',
    'OperatingSystem',
    'Organization',
    'Person',
    'PhoneNumber',
    'Product',
    'ProgrammingLanguage',
    'ProvinceOrState',
    'PublishedMedium',
    'RadioProgram',
    'RadioStation',
    'Region',
    'SportsEvent',
    'SportsGame',
    'SportsLeague',
    'Technology',
    'TVShow',
    'TVStation',
    'URL',
    'CalaisDocumentCategory', // Special reserved vocab for Calais Document Categorization
    'EventsFacts', // Special reserved vocab for Events & Facts
  );
}

//////////////////////////////////////////////////////////////////////////////
// ARC2 interoperability
// Copied from rdf.api.inc to fix bug http://drupal.org/node/348758
// When that is resolved and in a release, remove this hackitude.

function _calais_api_deconstruct_arc2_triple($triple) {
  $subject   = _calais_api_deconstruct_arc2_value($triple, 's');
  $predicate = _calais_api_deconstruct_arc2_value($triple, 'p');
  $object    = _calais_api_deconstruct_arc2_value($triple, 'o');
  return array($subject, $predicate, $object);
}

function _calais_api_deconstruct_arc2_value($triple, $name) {
  switch ($name == 'p' ? 'iri' : $triple[$name .'_type']) {
    case 'uri': // THIS IS THE FIX
    case 'iri':
      return rdf_uri($triple[$name]);
    case 'bnode':
      return rdf_bnode($triple[$name]);
    case 'literal':
    case 'literal1':
    case 'literal2':
    case 'literal_long1':
    case 'literal_long2':
      return rdf_literal($triple[$name], $triple[$name .'_lang'], $triple[$name .'_datatype']);
    case 'var':
      return rdf_var($triple[$name]);
  }
}

/**
 * This function exists as long as http://drupal.org/node/350485 does.
 *
 * This makes sure all of the extra Calais namespaces are in there. The original
 * RDF from Calais did not use namespaces as everything was an rdf:Description with
 * a type, but the RDF module serialization moves that type to the root tag level
 * and as such additional namespaces are needed.
 */
function _calais_api_fix_rdf_namespaces($xml) {
  $matches = array();
  preg_match('/<rdf:RDF (.*)>/i', $xml, $matches);
  $ns = explode(' ', $matches[1]);
  
  $calaisns = calais_api_rdf_namespaces();
  // Remove already specified namespaces from our list of those to add so we dont duplicate
  foreach($ns as $index => $namespace) {
    list($key, $value) = explode('=', $namespace); // get the attribute name
    list($xmlns, $prefix) = explode(':', $key); // strip the xmlns: off the front
    unset($calaisns[$prefix]);
  }
  
  // Add the namespaces that are left
  foreach ($calaisns as $prefix => $uri) {
    $ns[] = "xmlns:$prefix=\"$uri\"";
  }
  $namespaces = implode(' ', $ns);
  $rdf = preg_replace('/<rdf:RDF .*>/i', "<rdf:RDF $namespaces>", $xml, 1);
  return $rdf;
}