<?php $images = &get_children('post_type=attachment&post_mime_type=image&orderby=menu_order&order=ASC&post_parent='.$post->ID);
    if ($images):
    $image = current($images); $imageData = wp_get_attachment_image_src($image->ID); ?>
    <img src="<?php echo $imageData[0] ?>" alt="<?php the_title() ?>" />
    <?php endif ?>
<!-- image title: $image->post_title, caption: $image->post_excerpt, description: $image->post_content -->

Usually we’d use this just to pull out the first image attachment to a post and use this in a special place in our template. Now that there is built-in post thumbnail support, we wanted to use this feature but also still have the ability to display the selected thumbnail’s caption information. the_post_thumbnail only returns the image tag, so here is our custom filter (inside our template’s functions.php file) that adds our caption:

function monahans_thumbnail_caption($html, $post_id, $post_thumbnail_id, $size, $attr)
{
  $attachment =& get_post($post_thumbnail_id);
 
  // post_title => image title
  // post_excerpt => image caption
  // post_content => image description
 
  if ($attachment->post_excerpt || $attachment->post_content) {
    $html .= '<p class="thumbcaption">';
    if ($attachment->post_excerpt) {
      $html .= '<span class="captitle">'.$attachment->post_excerpt.'</span> ';
    }
    $html .= $attachment->post_content.'</p>';
  }
 
	return $html;
}
 
add_action('post_thumbnail_html', 'monahans_thumbnail_caption', null, 5);

The result of this filter is now anytime we call the_post_thumbnail in our template, the html will come back with our image caption, too.

http://www.php.net/manual/en/function.http-build-query.php

The way symfony parses its configuration yml files, it does a deep merge and makes the settings for “all” override the settings for “default”. So if you had an “all” setting in your application’s security.yml, and then a “default” setting in your module, the “all” setting would still take precedence even though mentally we like to think of module settings always overriding what is found in the application-wide files.

Here is the snippet in of code that actually does this processing, (it’s in a file called sfSecurityConfigHandler.class.php). Here, $myConfig is an array with action names as the keys, and the value is an array containing the “is_secure” parameter along with any specific credentials required.

$myConfig['all'] = sfToolkit::arrayDeepMerge(
      isset($myConfig['default']) && is_array($myConfig['default']) ? $myConfig['default'] : array(),
      isset($myConfig['all']) && is_array($myConfig['all']) ? $myConfig['all'] : array()
    );

So, when symfony is parsing all of those security.yml files, you end up with a bunch of entries for all of your various actions:

$myConfig['login'] = array('is_secure' => false);
$myConfig['mySecureAction'] = array('is_secure' => true, 'credentials' => array(...));

and then symfony creates the “all” setting, based on what it finds for “default” and “all”, which it will use for any action that doesn’t have an explicit security setting. This $myConfig array is what eventually gets saved into the auto-generated [module]_config_security.yml.php files inside your cache directory.

So, if you are having trouble getting the right javascript or stylesheets to load in your view, or having trouble securing a page even though you are certain it’s listed in your security.yml file, check to see if there is an existing “all” configuration that might be overriding your local “default”.

PHPExcel

I’ve only been working with it for a few hours but it’s amazing. It can read and write from many different formats.

Output your spreadsheet object to different file formats

• Excel 2007 (spreadsheetML)
• BIFF8 (Excel 97 and higher)
• PHPExcel Serialized Spreadsheet
• CSV (Comma Separated Values)
• HTML
• PDF

Read different file formats into your spreadsheet object

• Excel 2007 (spreadsheetML)
• BIFF5 (Excel 5.0 / Excel 95), BIFF8 (Excel 97 and higher)
• PHPExcel Serialized Spreadsheet
• CSV (Comma Separated Values)

I am very excited about this. Another honorable mention is php-excel-reader which might be worth a look if you just want something a bit more simple.

Let’s say you want to add a custom column to your query, let’s say “points remaining” for the sake of example.

  $c->addAsColumn('points_remaining', '('.UserPointsPeer::TOTAL_POINTS.' - '.UserPointsPeer::POINTS_EARNED.')');

And now you want to limit your query to just users who have a certain number of points remaining. Since this is on a custom column, you have to use the addHaving() method. This gets a bit more tricky because when adding criteria based on custom columns, Propel wants to generate criterion objects using column names which include the column’s table name and dot separator. So, we have to rely on the Criteria::CUSTOM functionality. No doubt this is a hack and it completely circumvents any built-in query escaping that is essential, but if you are careful about sanitizing your input, you can do the following. Note this code sample assumes that $pointsFilter['from'] and $pointsFilter['to'] contain the GET or POST input parameters for use in your query:

// first version where we want to fetch a custom column
if (!empty($pointsFilter['from'])) {
  $pointsFilter['from'] = (int) $pointsFilter['from'];
  $criterion = $c->getNewCriterion(UserPointsPeer::ID, 'points_remaining >= '.$pointsFilter['from'], Criteria::CUSTOM);
}
 
if (!empty($pointsFilter['to'])) {
  $pointsFilter['to'] = (int) $pointsFilter['to'];
  if (isset($criterion)) {
    $criterion->addAnd($c->getNewCriterion(UserPointsPeer::ID, 'points_remaining <= '.$pointsFilter['to'], Criteria::CUSTOM));
  } else {
    $criterion = $c->getNewCriterion(UserPointsPeer::ID, 'points_remaining <= '.$pointsFilter['to'], Criteria::CUSTOM);
  }
}
 
if (isset($criterion)) {
  $c->addHaving($criterion);
}

Note that we have to use addHaving() to add our condition to our custom column.

If you don’t need to retrieve the custom column in your query but still want to limit your results on that criteria, you can use the add() method, but you have to change it up a bit:

// second version without adding a custom column
if (!empty($pointsFilter['from'])) {
  $pointsFilter['from'] = (int) $pointsFilter['from'];
  $criterion = $c->getNewCriterion(
    UserPointsPeer::ID, // can be any column in your table, really
    '('.CurriculumPeer::TOTAL_POINTS.' - '.UserPointsPeer::POINTS_EARNED.') >= '.$pointsFilter['from'], 
    Criteria::CUSTOM
  );
}
 
if (!empty($pointsFilter['to'])) {
  $pointsFilter['to'] = (int) $pointsFilter['to'];
  $pointsFilterToCriterion = $c->getNewCriterion(
    UserPointsPeer::ID, 
    '('.CurriculumPeer::TOTAL_POINTS.' - '.UserPointsPeer::POINTS_EARNED.') <= '.$pointsFilter['to'],
    Criteria::CUSTOM
  );
 
  if (isset($criterion)) {
    $criterion->addAnd($pointsFilterToCriterion);
  } else {
    $criterion = $pointsFilterToCriterion;
  }
}
 
if (isset($criterion)) {
  $c->add($criterion);
}

Note that in this example I’m typecasting our input parameters are integers, but you could mysql_real_escape_string() or any other sanitizing method you like.

To use an example off of the doctrine-user group, imagine you have a Content object and this may or may not have a related Metadata record.

If you are familiar with Propel, then if you had your hydrated $content object, you could do

// with propel:
if ($metadata = $content->getMetadata()) {
  // do something with the related $metadata
}

and getMetadata() would return NULL if there was no related record to be found.

Doctrine will automatically create an empty related object if you request one, so calling $content->Metadata or $content['Metadata'] or $content->getMetadata() will always return a Metadata object, even if it does not exist in the database. So, you can’t rely on the same trick above to see if a related object exists.

So, how do you check if a related object actually exists? The first way, and my favorite, is using the exists() method to check if the related object is persistent (i.e. the object is saved in the DB).

// with doctrine 
if ($content->Metadata->exists()) { 
  // related metadata exists for this $content object
}

Using isset() will check if the related object has been hydrated/ loaded, but it will not do a query against the database to see if it exists there. So, if you fetch a Content object and do not do any joins in your query to join the related Metadata objects, then isset(), either by doing isset($business->Metadata) or $business->isset('Metadata') will *always* return false.

If your foreign key to the related object exists in your parent record (if content has a metadata_id field) you can always check the value of $content['metadata_id']. This will not work if the foreign key exists only in your metadata table (metadata has a column content_id, but content does not have a metadata_id column).

I think if you really wanted to do it right, if you expect to be checking for the existence of a related object (“Metadata”) on your collection of core objects (“Content”), you should do a join in your query so you don’t have to hit the database for every single Content record fetched to see if a related Metadata record exists.

Here are a few threads on this subject in the doctrine-user group:

• Call $record->Related without creating new object
• hasOne relationship
• Awkward handling of One-to-Zero-or-One relationships in Doctrine

if (isset($myarray['foo']) && $myarray['foo'])) {
  // do something
}

But I thought there had to be a better way, and one that would not throw undefined errors. A quick google search came up with a few good hits, include the PHP manual page for empty().

The manual says that when using empty(), no warnings are thrown when a variable is not set. It will return true in the following cases:

“” (an empty string)
0 (0 as an integer)
“0″ (0 as a string)
NULL
FALSE
array() (an empty array)
var $var; (a variable declared, but without a value in a class)

Keep in mind that an object with no properties is not considered empty().

Now, we can make our code a bit more beautiful:

if (!empty($myarray['foo'])) {
  // do something
}

I’ve known about empty() for a long time, but for some reason always seemed to choose isset() over empty(), even in cases like this.

1) Use another table to list the enum options. Pros: most flexible. Works nicely within symfony. Cons: means more joins in your queries.

2) Use enum anyway. You can add custom column types using a little known trick in your schema:

status:
  id:
  name: { type: varchar, sqltype:enum, size: "'inactive','active'", default: inactive }

Pros: You get your enum. Cons: Not very portable, a bit of a hack.

3) Simulate an enum column using your model. I typically end up doing this for columns that have a finite number of defined options *that will not change*. Pros: Fast. Cons: Not as flexible as optino 1. Example:

class UserProfilePeer extends BaseUserProfilePeer
{
  static protected $USER_TYPE_INTEGERS = array('student', 'supervisor', 'coordinator', 'researcher', 'admin');
  static protected $USER_TYPE_VALUES = null;

  /**
   * Returns a user type label from an index value.
   *
   * @param integer $index
   * @return string user type label
   * @author Scott Meves
   */
  static public function getUserTypeFromIndex($index)
  {
    return self::$USER_TYPE_INTEGERS[$index];
  }

  /**
   * Returns the user type index from a string value
   *
   * @param string $value
   * @return integer $index
   * @author Scott Meves
   */
  static public function getUserTypeFromValue($value)
  {
    if (!self::$USER_TYPE_VALUES)
    {
      self::$USER_TYPE_VALUES = array_flip(self::$USER_TYPE_INTEGERS);
    }

    $values = strtolower($value);

    if (!isset(self::$USER_TYPE_VALUES[$value]))
    {
      throw new PropelException(sprintf('User type cannot take "%s" as a value', $value));
    }

    return self::$USER_TYPE_VALUES[strtolower($value)];
  }

  public static function getUserTypeOptions()
  {
    return self::$USER_TYPE_INTEGERS;
  }
}

Oftentimes I find myself wanting to return the contents of a partial as the entire response for an ajax request. I do this a lot when I am already using a partial in my template, and then I want to replace that section with the results of an ajax call. We already have a partial set up to display that content, so we might as well reuse it when we have to refresh that area with new content.

Imagine we have a page that presents a list of articles. In the right sidebar we’d like to feature a preview of the first article. We’d also like to allow a “quick look” for the other articles. Clicking on the quick look button should load a preview of the selected article in the right sidebar, replaced the default preview.

To do this, I set up a partial that contains the article preview. In the default list template (e.g. listSuccess.php) I include the partial along with the featured article.

<?php include_partial('preview', array('article'=>$article)) ?>

Then, when someone clicks the quick look button, we want to make an ajax call that returns the new preview content with that particular article. Our action might look something like this:

public function executeLoadPreview()
{
  $this->article = ArticlePeer::retrieveByPk($this->getRequestParamer('articleId'));
}

The most obvious thing to do now would be to set up a standard loadPreviewSuccess.php tempate and within that template simply include the partial just as you would normally.

// loadPreviewSuccess.php
<?php include_partial('preview', array('article'=>$article) ?>

This is totally fine, and you may like it, but there is another way to do it. We can just render the contents of the partial directly from our action. Here is what it looks like. (Note that I turn off the web debugging so that it doesn’t render within the ajax call!)

public function executeLoadPreview()
{
  sfConfig::set('sf_web_debug', false);
  sfLoader::loadHelpers('Partial');
 
  $article = ArticlePeer::retrieveByPk($this->getRequestParamer('articleId'));
 
  return $this->renderText(get_partial('preview', array('article' => $article)));
}

Nice!

Let’s say you would like to allow users to upload PDF files to your site. You store data related to these file uploads in an “article” table in your database. This table has a column named “file” that stores the file name of the uploaded pdf.

The symfony forms framework will generate the basic (and not so basic) code to get you started. If you haven’t yet generated your form classes yet, you can do so with the command:

./symfony propel:build-forms

Now navigate to lib/form/ArticleForm.class.php. To convert the file column to a file input field, we do the following:

class ArtistResumeForm extends BaseArtistResumeForm
{
  public function configure()
  {
    $this->widgetSchema['file']        = new sfWidgetFormInputFile();
 
    $this->validatorSchema['file']     = new sfValidatorFile(array('path' => sfConfig::get('sf_upload_dir').'/articles', 'required' => false));
  }
}

Now, whenever you call $form->save(), your files will automatically be uploaded to the server, and the field “file” in your propel object will be updated with the new file name so you can easily retrieve it later.

If you are interested in getting this to work with embedded forms, check out this other post on uploading files within embedded forms with symfony 1.2.