Better Facebook Titles: mod

[ghostofzigz]

UPDATED Jan 12, 2017 (working well)

My News Aggrigation application filters by title. The titles coming from the FacebookBridge are mostly NON-unique. Here is an update to the FacebookBridge title section which gives a far more descriptive title and attempts to make it unique if there is not enough data:

On or near line 150 find the following
AFTER THIS, add:

			// remove these items from title
			$_d1 = html_entity_decode("·", ENT_COMPAT, "UTF-8");
			$_d2 = html_entity_decode("Â", ENT_COMPAT, "UTF-8");
			$rp = array(
				"/$_d1/",
				"/$_d2/",
				"/^PostsSee All/",
				"/^NotesSee All/",
				"/$author shared/",
				"/^$author/",
				"/^shared/",
				"/^\&quot\;/",
				"/^Posts/",
				"/^Notes/",
				"/^Continue Reading/",
				"/^(\d*) hrs/",
				"/^Reviews/",
				"/^See All/",
				"/^(\d*) Reviews/",
				"/(.*?)See More/",
				"/^Click for more/",
			);

On or near Line 210 change:

					//Build title from username and content
					$title = $author;
					if (strlen($title) > 24)
						$title = substr($title, 0, strpos(wordwrap($title, 24), "\n")).'...';
					$title = $title.' | '.strip_tags($content);
					if (strlen($title) > 64)
						$title = substr($title, 0, strpos(wordwrap($title, 64), "\n")).'...';

To:

					//Build title from username and content
					/*
					$title = $author;
					if (strlen($title) > 24)
						$title = substr($title, 0, strpos(wordwrap($title, 24), "\n")).'...';
					$title = $title.' | '.strip_tags($content);
					if (strlen($title) > 64)
						$title = substr($title, 0, strpos(wordwrap($title, 64), "\n")).'...';

					*/
					
					// Build title from content, date, and username
					$p = "/.*?".$_d1."\s/";
					$title = strip_tags(preg_replace($p, '', $content));
					$_f = 1; $_c = 0;
					while ($_f) {
						$_f = 0;
						foreach ($rp as $p) {
							$title = trim(preg_replace($p, '', $title, $_c));
							if (($_c) && ($_c > 0)) $_f++;
						}
					}
					if (strlen($title) < 4) {
						$title = $author;
						if ($date) $_d = date("M j, Y", $date);
						$_l = (($date) ? 62 - strlen($_d) : 64);
						if (strlen($title) > $_l)
							$title = substr($title, 0, strpos(wordwrap($title, $_l), "\n"))."&hellip;";
						if ($date) $title = $title." ".$_d;
					} else {
						$_l = 64;
						if (strlen($title) > $_l)
							$title = substr($title, 0, strpos(wordwrap($title, $_l), "\n"))."&hellip;";
					}	

I also have a shell script to automate this update for whenever the RSS-Bridge code is updated from here... if anyone would like that, get at me.
-- ghostofzigz

[logmanoriginal]

Nice work @ghostofzigz 👍
Can you open a PR for this?

I haven't tested it on my machine yet, but can this be made compatible to all languages (I assume this will only work for English Facebook)?

Also, is there any particular reason to wrap lines at 24/64 characters or is it made to your specific needs?

[ghostofzigz]

I will take another look at simplifying the regex to eliminate language specific patterns... unfortunately at this time it looks like Facebook is blocking ALL requests via the RSS Bridge, so I will need to hack the request URL for this module to continue automated testing.

The reason for the 24 and 64 length limits was for compatibility with the code I was replacing (wasn't sure the reason for that either, figured it may come into play in other parts of the application adding to the title, and twitter length limits).

[logmanoriginal]

Closing because Facebook titles were fixed a long time ago.