You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

337 lines
9.9 KiB

8 years ago
8 years ago
  1. <?php
  2. namespace XRay\Formats;
  3. class Mf2 {
  4. public static function parse($mf2, $url, $http) {
  5. if($item = $mf2['items'][0]) {
  6. // If the first item is a feed, the page is a feed
  7. if(in_array('h-feed', $item['type'])) {
  8. return self::parseHFeed($mf2, $http);
  9. }
  10. // Check each top-level h-card, and if there is one that matches this URL, the page is an h-card
  11. foreach($mf2['items'] as $i) {
  12. if(in_array('h-card', $i['type'])
  13. and array_key_exists('url', $i['properties'])
  14. ) {
  15. $urls = $i['properties']['url'];
  16. $urls = array_map('\normalize_url', $urls);
  17. if(in_array($url, $urls)) {
  18. // TODO: check for children h-entrys (like tantek.com), or sibling h-entries (like aaronparecki.com)
  19. // and return the result as a feed instead
  20. return self::parseHCard($i, $http, $url);
  21. }
  22. }
  23. }
  24. // Otherwise check for an h-entry
  25. if(in_array('h-entry', $item['type'])) {
  26. return self::parseHEntry($mf2, $http);
  27. }
  28. }
  29. return false;
  30. }
  31. private static function parseHEntry($mf2, $http) {
  32. $data = [
  33. 'type' => 'entry',
  34. 'author' => [
  35. 'type' => 'card',
  36. 'name' => null,
  37. 'url' => null,
  38. 'photo' => null
  39. ]
  40. ];
  41. $item = $mf2['items'][0];
  42. // Single plaintext values
  43. $properties = ['url','published','summary','rsvp'];
  44. foreach($properties as $p) {
  45. if($v = self::getPlaintext($item, $p))
  46. $data[$p] = $v;
  47. }
  48. // Always arrays
  49. $properties = ['photo','video','syndication','in-reply-to','like-of','repost-of','category'];
  50. foreach($properties as $p) {
  51. if(array_key_exists($p, $item['properties'])) {
  52. $data[$p] = [];
  53. foreach($item['properties'][$p] as $v) {
  54. if(is_string($v))
  55. $data[$p][] = $v;
  56. elseif(is_array($v) and array_key_exists('value', $v))
  57. $data[$p][] = $v['value'];
  58. }
  59. }
  60. }
  61. // Determine if the name is distinct from the content
  62. $name = self::getPlaintext($item, 'name');
  63. $content = null;
  64. $textContent = null;
  65. $htmlContent = null;
  66. if(array_key_exists('content', $item['properties'])) {
  67. $content = $item['properties']['content'][0];
  68. if(is_string($content)) {
  69. $textContent = $content;
  70. } elseif(!is_string($content) && is_array($content) && array_key_exists('value', $content)) {
  71. if(array_key_exists('html', $content)) {
  72. $textContent = trim(strip_tags($content['html']));
  73. $htmlContent = trim($content['html']);
  74. } else {
  75. $textContent = trim($content['value']);
  76. }
  77. }
  78. // Trim ellipses from the name
  79. $name = preg_replace('/ ?(\.\.\.|…)$/', '', $name);
  80. // Remove all whitespace when checking equality
  81. $nameCompare = preg_replace('/\s/','',trim($name));
  82. $contentCompare = preg_replace('/\s/','',trim($textContent));
  83. // Check if the name is a prefix of the content
  84. if(strpos($contentCompare, $nameCompare) === 0) {
  85. $name = null;
  86. }
  87. }
  88. if($name) {
  89. $data['name'] = $name;
  90. }
  91. if($content) {
  92. $data['content'] = [
  93. 'text' => $textContent
  94. ];
  95. if($textContent != $htmlContent) {
  96. $data['content']['html'] = $htmlContent;
  97. }
  98. }
  99. $data['author'] = self::findAuthor($mf2, $item, $http);
  100. return $data;
  101. }
  102. private static function parseHFeed($mf2, $http) {
  103. $data = [
  104. 'type' => 'feed',
  105. 'author' => [
  106. 'type' => 'card',
  107. 'name' => null,
  108. 'url' => null,
  109. 'photo' => null
  110. ],
  111. 'items' => [],
  112. 'todo' => 'Not yet implemented. Please see https://github.com/aaronpk/XRay/issues/1'
  113. ];
  114. return $data;
  115. }
  116. private static function parseHCard($item, $http, $authorURL=false) {
  117. $data = [
  118. 'type' => 'card',
  119. 'name' => null,
  120. 'url' => null,
  121. 'photo' => null
  122. ];
  123. $properties = ['url','name','photo'];
  124. foreach($properties as $p) {
  125. if($p == 'url' && $authorURL) {
  126. // If there is a matching author URL, use that one
  127. $found = false;
  128. foreach($item['properties']['url'] as $url) {
  129. $url = \normalize_url($url);
  130. if($url == $authorURL) {
  131. $data['url'] = $url;
  132. $found = true;
  133. }
  134. }
  135. if(!$found) $data['url'] = $item['properties']['url'][0];
  136. } else if($v = self::getPlaintext($item, $p)) {
  137. $data[$p] = $v;
  138. }
  139. }
  140. return $data;
  141. }
  142. private static function findAuthor($mf2, $item, $http) {
  143. $author = [
  144. 'type' => 'card',
  145. 'name' => null,
  146. 'url' => null,
  147. 'photo' => null
  148. ];
  149. // Author Discovery
  150. // http://indiewebcamp.com/authorship
  151. $authorPage = false;
  152. if(array_key_exists('author', $item['properties'])) {
  153. // Check if any of the values of the author property are an h-card
  154. foreach($item['properties']['author'] as $a) {
  155. if(self::isHCard($a)) {
  156. // 5.1 "if it has an h-card, use it, exit."
  157. return self::parseHCard($a, $http);
  158. } elseif(is_string($a)) {
  159. if(self::isURL($a)) {
  160. // 5.2 "otherwise if author property is an http(s) URL, let the author-page have that URL"
  161. $authorPage = $a;
  162. } else {
  163. // 5.3 "otherwise use the author property as the author name, exit"
  164. // We can only set the name, no h-card or URL was found
  165. $author['name'] = self::getPlaintext($item, 'author');
  166. return $author;
  167. }
  168. } else {
  169. // This case is only hit when the author property is an mf2 object that is not an h-card
  170. $author['name'] = self::getPlaintext($item, 'author');
  171. return $author;
  172. }
  173. }
  174. }
  175. // 6. "if no author page was found" ... check for rel-author link
  176. if(!$authorPage) {
  177. if(isset($mf2['rels']) && isset($mf2['rels']['author']))
  178. $authorPage = $mf2['rels']['author'][0];
  179. }
  180. // 7. "if there is an author-page URL" ...
  181. if($authorPage) {
  182. // 7.1 "get the author-page from that URL and parse it for microformats2"
  183. $authorPageContents = self::getURL($authorPage, $http);
  184. if($authorPageContents) {
  185. foreach($authorPageContents['items'] as $i) {
  186. if(self::isHCard($i)) {
  187. // 7.2 "if author-page has 1+ h-card with url == uid == author-page's URL, then use first such h-card, exit."
  188. if(array_key_exists('url', $i['properties'])
  189. and in_array($authorPage, $i['properties']['url'])
  190. and array_key_exists('uid', $i['properties'])
  191. and in_array($authorPage, $i['properties']['uid'])
  192. ) {
  193. return self::parseHCard($i, $http, $authorPage);
  194. }
  195. // 7.3 "else if author-page has 1+ h-card with url property which matches the href of a rel-me link on the author-page"
  196. $relMeLinks = (isset($authorPageContents['rels']) && isset($authorPageContents['rels']['me'])) ? $authorPageContents['rels']['me'] : [];
  197. if(count($relMeLinks) > 0
  198. and array_key_exists('url', $i['properties'])
  199. and count(array_intersect($i['properties']['url'], $relMeLinks)) > 0
  200. ) {
  201. return self::parseHCard($i, $http, $authorPage);
  202. }
  203. }
  204. }
  205. }
  206. // 7.4 "if the h-entry's page has 1+ h-card with url == author-page URL, use first such h-card, exit."
  207. foreach($mf2['items'] as $i) {
  208. if(self::isHCard($i)) {
  209. if(array_key_exists('url', $i['properties'])
  210. and in_array($authorPage, $i['properties']['url'])
  211. ) {
  212. return self::parseHCard($i, $http);
  213. }
  214. }
  215. }
  216. }
  217. return $author;
  218. }
  219. private static function responseDisplayText($name, $summary, $content) {
  220. // Build a fake h-entry to pass to the comments parser
  221. $input = [
  222. 'type' => ['h-entry'],
  223. 'properties' => [
  224. 'name' => [trim($name)],
  225. 'summary' => [trim($summary)],
  226. 'content' => [trim($content)]
  227. ]
  228. ];
  229. if(!trim($name))
  230. unset($input['properties']['name']);
  231. if(!trim($summary))
  232. unset($input['properties']['summary']);
  233. $result = \IndieWeb\comments\parse($input, false, 1024, 4);
  234. return [
  235. 'name' => trim($result['name']),
  236. 'content' => $result['text']
  237. ];
  238. }
  239. private static function hasNumericKeys(array $arr) {
  240. foreach($arr as $key=>$val)
  241. if (is_numeric($key))
  242. return true;
  243. return false;
  244. }
  245. private static function isMicroformat($mf) {
  246. return is_array($mf)
  247. and !self::hasNumericKeys($mf)
  248. and !empty($mf['type'])
  249. and isset($mf['properties']);
  250. }
  251. private static function isHCard($mf) {
  252. return is_array($mf)
  253. and !empty($mf['type'])
  254. and is_array($mf['type'])
  255. and in_array('h-card', $mf['type']);
  256. }
  257. private static function isURL($string) {
  258. return preg_match('/^https?:\/\/.+\..+$/', $string);
  259. }
  260. // Given an array of microformats properties and a key name, return the plaintext value
  261. // at that property
  262. // e.g.
  263. // {"properties":{"published":["foo"]}} results in "foo"
  264. private static function getPlaintext($mf2, $k, $fallback=null) {
  265. if(!empty($mf2['properties'][$k]) and is_array($mf2['properties'][$k])) {
  266. // $mf2['properties'][$v] will always be an array since the input was from the mf2 parser
  267. $value = $mf2['properties'][$k][0];
  268. if(is_string($value)) {
  269. return $value;
  270. } elseif(self::isMicroformat($value) && array_key_exists('value', $value)) {
  271. return $value['value'];
  272. }
  273. }
  274. return $fallback;
  275. }
  276. private static function getURL($url, $http) {
  277. if(!$url) return null;
  278. // TODO: consider adding caching here
  279. $result = $http->get($url);
  280. if($result['error'] || !$result['body']) {
  281. return null;
  282. }
  283. return \mf2\Parse($result['body'], $url);
  284. }
  285. }