<?php

/**
 * @file
 * @todo Add file header description
 */


# This script can be run as a stand-alone script to sync all the features from chado to drupal
// Parameter f specifies the feature_id to sync
// -f 0 will sync all features

$arguments = getopt("f:t:");

if (isset($arguments['f']) and isset($arguments['t']) and $arguments['t'] == 'chado_feature') {
  $drupal_base_url = parse_url('http://www.example.com');
  $_SERVER['HTTP_HOST'] = $drupal_base_url['host'];
  $_SERVER['REQUEST_URI'] = $_SERVER['SCRIPT_NAME'] = $_SERVER['PHP_SELF'];
  $_SERVER['REMOTE_ADDR'] = NULL;
  $_SERVER['REQUEST_METHOD'] = NULL;

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

  $feature_id = $arguments['f'];

  if ($feature_id > 0) {
    tripal_feature_sync_feature($feature_id);
  }
  else {
    print "syncing all features...\n";
    tripal_feature_sync_features();
  }
}
/**
 *
 */
function tripal_feature_sync_form() {

  $form['description'] = array(
  '#type' => 'item',
  '#value' => t("Add feature types, optionally select an organism and " .
     "click the 'Sync all Features' button to create Drupal " .
     "content for features in chado. Only features of the types listed " .
     "below in the Feature Types box will be synced. You may limit the " .
     "features to be synced by a specific organism. Depending on the " .
     "number of features in the chado database this may take a long " .
     "time to complete. "),
  );

  $form['feature_types'] = array(
    '#title'       => t('Feature Types'),
    '#type'        => 'textarea',
    '#description' => t("Enter the names of the feature types to sync.  Pages for these feature " .
       "types will be created automatically for features that exist in the " .
       "chado database.  The names listed here should be spearated by " .
       "spaces or entered separately on new lines. The names must match " .
       "exactly (spelling and case) with terms in the sequence ontology"),
    '#required'    => TRUE,
    '#default_value' => variable_get('chado_sync_feature_types', 'gene contig'),
  );

  // get the list of organisms
  $sql = "SELECT * FROM {organism} ORDER BY genus, species";
  $orgs = tripal_organism_get_synced();
  $organisms[] = '';
  foreach ($orgs as $organism) {
    $organisms[$organism->organism_id] = "$organism->genus $organism->species ($organism->common_name)";
  }
  $form['organism_id'] = array(
    '#title'       => t('Organism'),
    '#type'        => t('select'),
    '#description' => t("Choose the organism for which features types set above will be synced. Only organisms which also have been synced will appear in this list."),
    '#options'     => $organisms,
  );


  $form['button'] = array(
    '#type' => 'submit',
    '#value' => t('Sync all Features'),
    '#weight' => 3,
  );

  return $form;
}
/**
 *
 */
function tripal_feature_sync_form_validate($form, &$form_state) {
  $organism_id   = $form_state['values']['organism_id'];
  $feature_types = $form_state['values']['feature_types'];

  // nothing to do
}
/**
 *
 */
function tripal_feature_sync_form_submit($form, &$form_state) {

  global $user;

  $organism_id   = $form_state['values']['organism_id'];
  $feature_types = $form_state['values']['feature_types'];

  $job_args = array(0, $organism_id, $feature_types);

  if ($organism_id) {
    $organism = tripal_core_chado_select('organism', array('genus', 'species'), array('organism_id' => $organism_id));
    $title = "Sync all features for " . $organism[0]->genus . " " . $organism[0]->species;
  }
  else {
    $title = t('Sync all features for all synced organisms');
  }

  variable_set('chado_sync_feature_types', $feature_types);

  tripal_add_job($title, 'tripal_feature', 'tripal_feature_sync_features', $job_args, $user->uid);
}
/**
 *  
 * @param $na 
 *   Tripal expects all jobs to have at least one argument. For this function
 *   we don't need any, so we have this dummy argument as a filler
 * @param $job_id
 */
function tripal_feature_set_urls($na = NULL, $job = NULL) {
  
  // begin the transaction
  db_query("BEGIN");
      
  print "\nNOTE: Setting of URLs is performed using a database transaction. \n" .
        "If the load fails or is terminated prematurely then the entire set of \n" .
        "new URLs will be rolled back and no changes will be made\n\n";
  
  // get the number of records we need to set URLs for
  $csql = "SELECT count(*) FROM {chado_feature}";
  $num_nodes = db_result(db_query($csql));
    
  // calculate the interval at which we will print an update on the screen
  $num_set = 0;
  $num_per_interval = 100;
  
  // prepate the statements which will quickly add url alias. Because these
  // are not Chado tables we must manually prepare them 
  $psql = "
    PREPARE del_url_alias_by_src (text) AS
    DELETE FROM {url_alias} WHERE src = \$1
  ";
  db_query($psql);
  $psql = "
    PREPARE ins_url_alias_nisrds (text, text) AS
    INSERT INTO url_alias (src, dst) VALUES (\$1, \$2)
  ";
  db_query($psql);
  
  // get the URL alias syntax string
  $url_alias = variable_get('chado_feature_url_string', '/feature/[genus]/[species]/[type]/[uniquename]'); 
  if (!$url_alias) {
    $url_alias = '/feature/[genus]/[species]/[type]/[uniquename]';
  } 
  $url_alias = preg_replace('/^\//', '', $url_alias); // remove any preceeding forward slash
  
  
  // get the list of features that have been synced
  $sql = "SELECT * FROM {chado_feature}";
  $nodes = db_query($sql);  
  while ($node = db_fetch_object($nodes)) {
   
    // get the URL alias
    $src = "node/$node->nid";
    $dst = tripal_feature_get_feature_url($node, $url_alias);
    if (!$dst) {
      db_query('DEALLOCATE "del_url_alias_by_src"');
      db_query('DEALLOCATE "ins_url_alias_nisrds"');
      db_query("ROLLBACK"); 
      return; 
    }    
    
    // if the src and dst is the same (the URL alias couldn't be set)
    // then skip to the next one. There's nothing we can do about this one.
    if ($src == $dst) {
      continue;
    }
    
    // remove any previous alias and then add the new one
    $success = db_query("EXECUTE del_url_alias_by_src('%s')", $src);    
    if (!$success) {
      db_query('DEALLOCATE "del_url_alias_by_src"');
      db_query('DEALLOCATE "ins_url_alias_nisrds"');
      db_query("ROLLBACK");
      watchdog('trp-seturl', "Failed Removing URL Alias: %src", array('%src' => $src), WATCHDOG_ERROR);
      return;
    }
    $success = db_query("EXECUTE ins_url_alias_nisrds('%s', '%s')", $src, $dst);
    if (!$success) {
      db_query('DEALLOCATE "del_url_alias_by_src"');
      db_query('DEALLOCATE "ins_url_alias_nisrds"');
      db_query("ROLLBACK");
      watchdog('trp-seturl', "Failed Adding URL Alias: %dst", array('%dst' => $dst), WATCHDOG_ERROR);
      return;
    }

    // update the job status every 1% features
    if ($job and $num_set % $num_per_interval == 0) {
      $percent = ($num_set / $num_nodes) * 100;
      tripal_job_set_progress($job, intval($percent));
      $percent = sprintf("%.2f", $percent);
      print "Setting URLs (" . $percent . "%). Memory: " . number_format(memory_get_usage()) . " bytes.\r";
      
    }
    $num_set++;
  }
  $percent = ($num_set / $num_nodes) * 100;
  tripal_job_set_progress($job, intval($percent));
  $percent = sprintf("%.2f", $percent);
  print "Setting URLs (" . $percent . "%). Memory: " . number_format(memory_get_usage()) . " bytes.\r";
  print "\nDone. Set " . number_format($num_set) . " URLs\n";
  
  // unprepare the statements
  db_query('DEALLOCATE "del_url_alias_by_src"');
  db_query('DEALLOCATE "ins_url_alias_nisrds"');
  
  db_query("COMMIT");
}
/**
 * 
 * @param $node
 *   A node object containing at least the feature_id and nid
 * @param $url_alias
 *   Optional.  This should be the URL alias syntax string that contains
 *   placeholders such as [id], [genus], [species], [name], [uniquename],
 *   and [type].  These placeholders will be substituted for actual values.
 *   If this parameter is not provided then the value of the 
 *   chado_feature_url_string Drupal variable will be used.
 */
function tripal_feature_get_feature_url($node, $url_alias = NULL) {

  // get the starting URL alias
  if (!$url_alias) {
    $url_alias = variable_get('chado_feature_url_string', '/feature/[genus]/[species]/[type]/[uniquename]'); 
    if (!$url_alias) {
      $url_alias = '/feature/[genus]/[species]/[type]/[uniquename]';
    } 
    $url_alias = preg_replace('/^\//', '', $url_alias); // remove any preceeding forward slash
  }

  // get the feature 
  $values = array('feature_id' => $node->feature_id); 
  $options = array('statement_name' => 'sel_feature_id');       
  $feature = tripal_core_chado_select('feature', array('*'), $values, $options);
  if (!$feature) {
    watchdog('trp-seturl', "Cannot find feature when setting URL alias for feature: %id", array('%id' => $node->feature_id), WATCHDOG_ERROR);
    return FALSE;  
  }
  $feature = (object) $feature[0];
  
  // get the organism
  $values = array('organism_id' => $feature->organism_id);
  $options = array('statement_name' => 'sel_organism_id');
  $organism  = tripal_core_chado_select('organism', array('*'), $values, $options);  
  if (!$organism) {
    watchdog('trp-seturl', "Cannot find organism when setting URL alias for feature: %id", array('%id' => $node->feature_id), WATCHDOG_ERROR);
    return FALSE;  
  }
  $genus = preg_replace('/\s/', '_', strtolower($organism[0]->genus));
  $species = preg_replace('/\s/', '_', strtolower($organism[0]->species)); 

  // get the type
  $values = array('cvterm_id' => $feature->type_id);
  $options = array('statement_name' => 'sel_cvterm_id');
  $cvterm = tripal_core_chado_select('cvterm', array('name'), $values, $options);
  if (!$cvterm) {
    watchdog('trp-seturl', "Cannot find type when setting URL alias for feature: %id", array('%id' => $node->feature_id), WATCHDOG_ERROR);
    return FALSE;  
  }
  $type = preg_replace('/\s/', '_', $cvterm[0]->name);
  
  // now substitute in the values
  $url_alias = preg_replace('/\[id\]/', $feature->feature_id, $url_alias);
  $url_alias = preg_replace('/\[genus\]/', $genus, $url_alias);
  $url_alias = preg_replace('/\[species\]/', $species, $url_alias);
  $url_alias = preg_replace('/\[type\]/', $type, $url_alias);
  $url_alias = preg_replace('/\[name\]/', $feature->name, $url_alias);
  $url_alias = preg_replace('/\[uniquename\]/', $feature->uniquename, $url_alias);
 
  // the dst field of the url_alias table is only 128 characters long. 
  // if this is the case then simply return the node URL, we can't set this one
  if (strlen($url_alias) > 128) {
    watchdog('trp-seturl', "Cannot set alias longer than 128 characters: %alias.", array('%alias' => $url_alias), WATCHDOG_ERROR);
    return "node/" . $node->nid;
  }
  
  return $url_alias;
}
/**
 *
 *
 * @ingroup tripal_feature
 */
function tripal_feature_sync_features($max_sync = 0, $organism_id = NULL,
  $feature_types = NULL, $job_id = NULL) {
  
  $i = 0;

  // get the list of available sequence ontology terms for which
  // we will build drupal pages from features in chado.  If a feature
  // is not one of the specified typse we won't build a node for it.
  if (!$feature_types) {
    $allowed_types = variable_get('chado_sync_feature_types', 'gene contig');
  }
  else {
    $allowed_types = $feature_types;
  }
  $allowed_types = preg_replace("/[\s\n\r]+/", " ", $allowed_types);

  print "Looking for features of type: $allowed_types\n";

  $so_terms = split(' ', $allowed_types);
  $where_cvt = "";
  foreach ($so_terms as $term) {
    $where_cvt .= "CVT.name = '$term' OR ";
  }
  $where_cvt = drupal_substr($where_cvt, 0, drupal_strlen($where_cvt)-3);  # strip trailing 'OR'

  // get the list of organisms that are synced and only include features from
  // those organisms
  $orgs = tripal_organism_get_synced();
  $where_org = "";
  foreach ($orgs as $org) {
    if ($organism_id) {
      if ($org->organism_id and $org->organism_id == $organism_id) {
        $where_org .= "F.organism_id = $org->organism_id OR ";
      }
    }
    else {
    if ($org->organism_id) {
      $where_org .= "F.organism_id = $org->organism_id OR ";
    }
    }
  }
  $where_org = drupal_substr($where_org, 0, drupal_strlen($where_org)-3);  # strip trailing 'OR'

  // use this SQL statement to get the features that we're going to upload
  $sql = "SELECT feature_id " .
        "FROM {FEATURE} F " .
        "  INNER JOIN {Cvterm} CVT ON F.type_id = CVT.cvterm_id " .
        "  INNER JOIN {CV} on CV.cv_id = CVT.cv_id " .
        "WHERE ($where_cvt) AND ($where_org) AND CV.name = 'sequence' " .
        "ORDER BY feature_id";

  // get the list of features
  $results = chado_query($sql);

  // load into ids array
  $count = 0;
  $ids = array();
  while ($id = db_fetch_object($results)) {
    $ids[$count] = $id->feature_id;
    $count++;
  }

  // make sure our vocabularies are set before proceeding
  tripal_feature_set_vocabulary();

  // pre-create the SQL statement that will be used to check
  // if a feature has already been synced.  We skip features
  // that have been synced
  $sql = "SELECT * FROM {chado_feature} WHERE feature_id = %d";

  // Iterate through features that need to be synced
  $interval = intval($count * 0.01);
  if ($interval < 1) {
    $interval = 1;
  }
  $num_ids = sizeof($ids);
  $i = 0;
  foreach ($ids as $feature_id) {
    // update the job status every 1% features
    if ($job_id and $i % $interval == 0) {
      tripal_job_set_progress($job_id, intval(($i/$count)*100));
    }
    // if we have a maximum number to sync then stop when we get there
    // if not then just continue on
    if ($max_sync and $i == $max_sync) {
      return '';
    }
    if (!db_fetch_object(db_query($sql, $feature_id))) {

      # parsing all the features can cause memory overruns
      # we are not sure why PHP does not clean up the memory as it goes
      # to avoid this problem we will call this script through an
      # independent system call
      print "$i of $num_ids Syncing feature id: $feature_id\n";
      $cmd = "php " . drupal_get_path('module', 'tripal_feature') . "/includes/tripal_feature.sync_features.inc -f $feature_id -t chado_feature";
      system($cmd);

    }
    $i++;
  }

  return '';
}

/**
 *
 *
 * @ingroup tripal_feature
 */
function tripal_feature_sync_feature($feature_id) {
  //print "\tSyncing feature $feature_id\n";

  global $user;
  $create_node = 1;   // set to 0 if the node exists and we just sync and not create

  // get the accession prefix
  $aprefix = variable_get('chado_feature_accession_prefix', 'FID');

  // if we don't have a feature_id then return
  if (!$feature_id) {
    drupal_set_message(t("Please provide a feature_id to sync"));
    return '';
  }

  // get information about this feature
  $fsql = "SELECT F.feature_id, F.name, F.uniquename,O.genus, " .
         "    O.species,CVT.name as cvname,F.residues,F.organism_id " .
         "FROM {FEATURE} F " .
         "  INNER JOIN {Cvterm} CVT ON F.type_id = CVT.cvterm_id " .
         "  INNER JOIN {Organism} O ON F.organism_id = O.organism_ID " .
         "WHERE F.feature_id = %d";
  $feature = db_fetch_object(chado_query($fsql, $feature_id));

  // get the synonyms for this feature
  $synsql = "SELECT S.name " .
            "FROM {feature_synonym} FS " .
            "  INNER JOIN {synonym} S on FS.synonym_id = S.synonym_id " .
            "WHERE FS.feature_id = %d";
  $synonyms = chado_query($synsql, $feature_id);

  // now add these synonyms to the feature object as a single string
  $synstring = '';
  while ($synonym = db_fetch_object($synonyms)) {
    $synstring .= "$synonym->name\n";
  }
  $feature->synonyms = $synstring;

  // check to make sure that we don't have any nodes with this feature name as a title
  // but without a corresponding entry in the chado_feature table if so then we want to
  // clean up that node.  (If a node is found we don't know if it belongs to our feature or
  // not since features can have the same name/title.)
  $tsql =  "SELECT * FROM {node} N " .
           "WHERE title = '%s'";
  $cnsql = "SELECT * FROM {chado_feature} " .
           "WHERE nid = %d";
  $nodes = db_query($tsql, $feature->name);
  // cycle through all nodes that may have this title
  while ($node = db_fetch_object($nodes)) {
    $feature_nid = db_fetch_object(db_query($cnsql, $node->nid));
    if (!$feature_nid) {
      drupal_set_message(t("%feature_id: A node is present but the chado_feature entry is missing... correcting", array('%feature_id' => $feature_id)));
      node_delete($node->nid);
    }
  }

  // check if this feature already exists in the chado_feature table.
  // if we have a chado feature, we want to check to see if we have a node
  $cfsql = "SELECT * FROM {chado_feature} " .
           "WHERE feature_id = %d";
  // @coder-ignore: don't need to use db_rewrite_sql() since need all nodes regardless of access control
  $nsql =  "SELECT * FROM {node} N " .
           "WHERE nid = %d";
  $chado_feature = db_fetch_object(db_query($cfsql, $feature->feature_id));
  if ($chado_feature) {
    drupal_set_message(t("%feature_id: A chado_feature entry exists", array('%feature_id' => $feature_id)));
    $node = db_fetch_object(db_query($nsql, $chado_feature->nid));
    if (!$node) {
      // if we have a chado_feature but not a node then we have a problem and
      // need to cleanup
      drupal_set_message(t("%feature_id: The node is missing, but has a chado_feature entry... correcting", array('%feature_id' => $feature_id)));
      $df_sql = "DELETE FROM {chado_feature} WHERE feature_id = %d";
      db_query($df_sql, $feature_id);
    }
    else {
      drupal_set_message(t("%feature_id: A corresponding node exists", array('%feature_id' => $feature_id)));
      $create_node = 0;
    }
  }

  // if we've encountered an error then just return.
  if ($error_msg = db_error()) {
    //print "$error_msg\n";
    return '';
  }

  // if a drupal node does not exist for this feature then we want to
  // create one.  Note that the node_save call in this block
  // will call the hook_submit function which
  if ($create_node) {
    // get the organism for this feature
    $sql = "SELECT * FROM {organism} WHERE organism_id = %d";
    $organism = db_fetch_object(chado_query($sql, $feature->organism_id));

    drupal_set_message(t("%feature_id: Creating node $feature->name", array('%feature_id' => $feature_id)));
    $new_node = new stdClass();
    $new_node->type = 'chado_feature';
    $new_node->uid = $user->uid;
    $new_node->title = "$feature->name, $feature->uniquename ($feature->cvname) $organism->genus $organism->species";
    $new_node->fname = "$feature->name";
    $new_node->uniquename = "$feature->uniquename";
    $new_node->feature_id = $feature->feature_id;
    $new_node->residues = $feature->residues;
    $new_node->organism_id = $feature->organism_id;
    $new_node->feature_type = $feature->cvname;
    $new_node->synonyms = $feature->synonyms;

    // validate the node and if okay then submit
    node_validate($new_node);
    if ($errors = form_get_errors()) {
      print "Error encountered validating new node. Cannot sync\n";
      foreach ($errors as $key => $msg) {        
        watchdog('trp-fsync', "%msg", array('%msg' => $msg), 'error');
      }
      exit;
    }
    else {
      $node = node_submit($new_node);
      node_save($node);
    }
  }
  else {
    $node = $chado_feature;
  }


  // set the taxonomy for this node
  drupal_set_message(t("%feature_id ($node->nid): setting taxonomy", array('%feature_id' => $feature_id)));
  tripal_feature_set_taxonomy($node, $feature_id);


  return '';
}