X-Git-Url: https://git.immae.eu/?a=blobdiff_plain;f=inc%2Fpoche%2FUrl.class.php;h=5a8930144358674169bf94746aad80a8f725226c;hb=ce4a1dcc193d3124fc51c4942137621d3d806a19;hp=36fdd0f5b445ab27daa6c40b96a9070cd33921ff;hpb=8340fedd6403e5ef68df443476a548ff0d241b67;p=github%2Fwallabag%2Fwallabag.git
diff --git a/inc/poche/Url.class.php b/inc/poche/Url.class.php
index 36fdd0f5..5a893014 100644
--- a/inc/poche/Url.class.php
+++ b/inc/poche/Url.class.php
@@ -12,6 +12,45 @@ class Url
{
public $url;
+ private $fingerprints = array(
+ // Posterous
+ ' array('hostname'=>'fingerprint.posterous.com', 'head'=>true),
+ // Blogger
+ ' array('hostname'=>'fingerprint.blogspot.com', 'head'=>true),
+ ' array('hostname'=>'fingerprint.blogspot.com', 'head'=>true),
+ // WordPress (self-hosted and hosted)
+ '";
+ } else {
+ $html = "Download $_name";
+ }
+ $title = $_name;
+ $do_content_extraction = false;
+ break;
+ }
+ }
+ }
+ unset($_mime, $_act, $_name, $match);
+ }
+ if ($do_content_extraction) {
+ $html = $response['body'];
+ // remove strange things
+ $html = str_replace('[>', '', $html);
+ $html = $this->convert_to_utf8($html, $response['headers']);
+
+ // check site config for single page URL - fetch it if found
+ if ($single_page_response = $this->getSinglePage($item, $html, $effective_url)) {
+ $html = $single_page_response['body'];
+ // remove strange things
+ $html = str_replace('[>', '', $html);
+ $html = $this->convert_to_utf8($html, $single_page_response['headers']);
+ $effective_url = $single_page_response['effective_url'];
+ unset($single_page_response);
+ }
+ $extract_result = $extractor->process($html, $effective_url);
+ $readability = $extractor->readability;
+ $content_block = ($extract_result) ? $extractor->getContent() : null;
+ }
+ }
+ if ($do_content_extraction) {
+ // if we failed to extract content...
+ if (!$extract_result) {
+ $html = $this->error_message;
+ // keep the original item description
+ $html .= $item->get_description();
+ } else {
+ $readability->clean($content_block, 'select');
+ if ($this->rewrite_relative_urls) $this->makeAbsolute($effective_url, $content_block);
+ if ($content_block->childNodes->length == 1 && $content_block->firstChild->nodeType === XML_ELEMENT_NODE) {
+ $html = $content_block->firstChild->innerHTML;
+ } else {
+ $html = $content_block->innerHTML;
+ }
+ // post-processing cleanup
+ $html = preg_replace('!
[\s\h\v]*
!u', '', $html); + } + } + } + + $title = ($extractor->getTitle() != '' ? $extractor->getTitle() : _('Untitled')); + $content = array ('title' => $title, 'body' => $html); + + return $content; } - public function fetchContent() + private function convert_to_utf8($html, $header=null) { - if ($this->isCorrect()) { - $this->clean(); - $html = Encoding::toUTF8(Tools::getFile($this->getUrl())); - - # if Tools::getFile() if not able to retrieve HTTPS content, try the same URL with HTTP protocol - if (!preg_match('!^https?://!i', $this->getUrl()) && (!isset($html) || strlen($html) <= 0)) { - $this->setUrl('http://' . $this->getUrl()); - $html = Encoding::toUTF8(Tools::getFile($this->getUrl())); + $encoding = null; + if ($html || $header) { + if (is_array($header)) $header = implode("\n", $header); + if (!$header || !preg_match_all('/^Content-Type:\s+([^;]+)(?:;\s*charset=["\']?([^;"\'\n]*))?/im', $header, $match, PREG_SET_ORDER)) { + // error parsing the response + } else { + $match = end($match); // get last matched element (in case of redirects) + if (isset($match[2])) $encoding = trim($match[2], "\"' \r\n\0\x0B\t"); + } + // TODO: check to see if encoding is supported (can we convert it?) + // If it's not, result will be empty string. + // For now we'll check for invalid encoding types returned by some sites, e.g. 'none' + // Problem URL: http://facta.co.jp/blog/archives/20111026001026.html + if (!$encoding || $encoding == 'none') { + // search for encoding in HTML - only look at the first 35000 characters + $html_head = substr($html, 0, 40000); + if (preg_match('/^<\?xml\s+version=(?:"[^"]*"|\'[^\']*\')\s+encoding=("[^"]*"|\'[^\']*\')/s', $html_head, $match)) { + $encoding = trim($match[1], '"\''); + } elseif (preg_match('/]+)/i', $html_head, $match)) { + $encoding = trim($match[1]); + } elseif (preg_match_all('/]+)>/i', $html_head, $match)) { + foreach ($match[1] as $_test) { + if (preg_match('/charset=["\']?([^"\']+)/i', $_test, $_m)) { + $encoding = trim($_m[1]); + break; + } + } + } + } + if (isset($encoding)) $encoding = trim($encoding); + // trim is important here! + if (!$encoding || (strtolower($encoding) == 'iso-8859-1')) { + // replace MS Word smart qutoes + $trans = array(); + $trans[chr(130)] = '‚'; // Single Low-9 Quotation Mark + $trans[chr(131)] = 'ƒ'; // Latin Small Letter F With Hook + $trans[chr(132)] = '„'; // Double Low-9 Quotation Mark + $trans[chr(133)] = '…'; // Horizontal Ellipsis + $trans[chr(134)] = '†'; // Dagger + $trans[chr(135)] = '‡'; // Double Dagger + $trans[chr(136)] = 'ˆ'; // Modifier Letter Circumflex Accent + $trans[chr(137)] = '‰'; // Per Mille Sign + $trans[chr(138)] = 'Š'; // Latin Capital Letter S With Caron + $trans[chr(139)] = '‹'; // Single Left-Pointing Angle Quotation Mark + $trans[chr(140)] = 'Œ'; // Latin Capital Ligature OE + $trans[chr(145)] = '‘'; // Left Single Quotation Mark + $trans[chr(146)] = '’'; // Right Single Quotation Mark + $trans[chr(147)] = '“'; // Left Double Quotation Mark + $trans[chr(148)] = '”'; // Right Double Quotation Mark + $trans[chr(149)] = '•'; // Bullet + $trans[chr(150)] = '–'; // En Dash + $trans[chr(151)] = '—'; // Em Dash + $trans[chr(152)] = '˜'; // Small Tilde + $trans[chr(153)] = '™'; // Trade Mark Sign + $trans[chr(154)] = 'š'; // Latin Small Letter S With Caron + $trans[chr(155)] = '›'; // Single Right-Pointing Angle Quotation Mark + $trans[chr(156)] = 'œ'; // Latin Small Ligature OE + $trans[chr(159)] = 'Ÿ'; // Latin Capital Letter Y With Diaeresis + $html = strtr($html, $trans); + } + if (!$encoding) { + $encoding = 'utf-8'; + } else { + if (strtolower($encoding) != 'utf-8') { + $html = SimplePie_Misc::change_encoding($html, $encoding, 'utf-8'); + /* + if (function_exists('iconv')) { + // iconv appears to handle certain character encodings better than mb_convert_encoding + $html = iconv($encoding, 'utf-8', $html); + } else { + $html = mb_convert_encoding($html, 'utf-8', $encoding); + } + */ + } } + } + return $html; + } + + private function makeAbsolute($base, $elem) { + $base = new SimplePie_IRI($base); + // remove '//' in URL path (used to prevent URLs from resolving properly) + // TODO: check if this is still the case + if (isset($base->path)) $base->path = preg_replace('!//+!', '/', $base->path); + foreach(array('a'=>'href', 'img'=>'src') as $tag => $attr) { + $elems = $elem->getElementsByTagName($tag); + for ($i = $elems->length-1; $i >= 0; $i--) { + $e = $elems->item($i); + //$e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e); + $this->makeAbsoluteAttr($base, $e, $attr); + } + if (strtolower($elem->tagName) == $tag) $this->makeAbsoluteAttr($base, $elem, $attr); + } + } - if (function_exists('tidy_parse_string')) { - $tidy = tidy_parse_string($html, array(), 'UTF8'); - $tidy->cleanRepair(); - - //Warning: tidy might fail so, ensure there is still a content - $body = $tidy->body(); - - //hasChildren does not seem to work, just check the string - //returned (and do not forget to clean the white spaces) - if (preg_replace('/\s+/', '', $body->value) !== "") - $html = $tidy->value; - } - - $parameters = array(); - if (isset($html) and strlen($html) > 0) - { - $readability = new Readability($html, $this->getUrl()); - $readability->convertLinksToFootnotes = CONVERT_LINKS_FOOTNOTES; - $readability->revertForcedParagraphElements = REVERT_FORCED_PARAGRAPH_ELEMENTS; - - if($readability->init()) - { - $content = $readability->articleContent->innerHTML; - $parameters['title'] = ($readability->articleTitle->innerHTML != '' ? $readability->articleTitle->innerHTML : _('Untitled')); - $parameters['content'] = $content; - - return $parameters; + private function makeAbsoluteAttr($base, $e, $attr) { + if ($e->hasAttribute($attr)) { + // Trim leading and trailing white space. I don't really like this but + // unfortunately it does appear on some sites. e.g.