X-Git-Url: http://git.shiar.net/minimedit.git/blobdiff_plain/f0a0679f7ec71d6ae4969d39da4d6be84147fcb4..85271527e754ad247974966b8899c1c2a3226b07:/article.inc.php

diff --git a/article.inc.php b/article.inc.php
index e3a3ee8..3c65a97 100644
--- a/article.inc.php
+++ b/article.inc.php
@@ -17,6 +17,7 @@ function showdate($parts)
 class ArchiveArticle
 {
 	public $raw, $preface, $title, $body;
+	public $meta = [];
 
 	function __construct($path)
 	{
@@ -24,6 +25,15 @@ class ArchiveArticle
 		$this->link = preg_replace('{(?:/index)?\.html$}', '', $path);
 		if (file_exists($this->page)) {
 			$this->raw = file_get_contents($this->page);
+
+			if (preg_match_all('{
+				\G <meta \s+ property="( [^"]+ )" \s+ content="( [^"]* )" > \s*
+			}x', $this->raw, $meta)) {
+				$matchlen = array_sum(array_map('strlen', $meta[0]));
+				$this->raw = substr($this->raw, $matchlen); # delete matched contents
+				$this->meta = array_combine($meta[1], $meta[2]); # [property => content]
+			}
+
 			@list ($this->preface, $this->title, $this->body) =
 				preg_split('{<h2>(.*?)</h2>\s*}', $this->raw, 2, PREG_SPLIT_DELIM_CAPTURE);
 		}
@@ -36,7 +46,7 @@ class ArchiveArticle
 
 	function safetitle()
 	{
-		return trim(strip_tags($this->title));
+		return trim($this->meta['og:title'] ?? strip_tags($this->title));
 	}
 	function name()
 	{
@@ -80,23 +90,29 @@ class ArchiveArticle
 		}
 		return $this->body;
 	}
+
 	function teaser()
 	{
-		if (preg_match('{
-			<meta \s+ name="description" [^>]* content="([^">]*)"
-		}x', $this->preface, $meta)) {
-			# prefer specific page description if found (assume before title)
-			#TODO: strip from body contents
-			return $meta[1];
+		if ($override = @$this->meta['og:description']) {
+			# prefer specific page description if found in metadata
+			return $override;
 		}
+
 		if (preg_match('{
-			</h2> (?: \s+ | <p\sclass="nav\b.*?</p> | <div[^>]*> )* <p> (.*?) </p>
+			</h2> (?: \s+ | <p\sclass="nav\b.*?</p> | <div[^>]*> )* <p> \s* (.*?) </p>
 		}sx', $this->raw, $bodyp, PREG_OFFSET_CAPTURE)) {
 			# fallback paragraph contents following the page header
 			if ($bodyp[1][1] < 256) {
 				return $bodyp[1][0];
 			}
 		}
+
+		# starting paragraph for documents without title (assumed simple/partial)
+		if (strpos($this->raw, '<h2') === FALSE and preg_match('{
+			\A <p> \s* (.*?) </p>
+		}sx', $this->raw, $bodyp)) {
+			return $bodyp[1];
+		}
 	}
 
 	function img()
@@ -107,6 +123,11 @@ class ArchiveArticle
 	}
 	function image()
 	{
+		if ($override = @$this->meta['og:image']) {
+			# prefer specific page image if found in metadata
+			return $override;
+		}
+
 		if ( preg_match('/\bsrc="([^"]*)"/', $this->img, $src) ) {
 			return $src[1];
 		}