/**
* This method handles running a test suite
* @param string $path: path to the suite's JSON file
* @access public
* @return bool
*/
public function runSuite($path)
{
$suite = json_decode(file_get_contents($path));
echo sprintf('Running %s.', $suite->name), "\n";
$iterator = new \DirectoryIterator(dirname($path));
# loop: each file in the test suite
foreach ($iterator as $file) {
# if: file is a sub-directory and not a dot-directory
if ($file->isDir() && !$file->isDot()) {
$this->tests_total++;
$path_of_test = $file->getPathname() . '/';
$test = json_decode(file_get_contents($path_of_test . 'test.json'));
$input = file_get_contents($path_of_test . 'input.html');
$expected_output = json_decode(file_get_contents($path_of_test . 'output.json'), TRUE);
$parser = new Parser($input, '', TRUE);
$output = $parser->parse(TRUE);
# if: test passed
if ($output['items'] === $expected_output['items']) {
// echo '.'; # can output a dot for successful tests
$this->tests_passed++;
} else {
echo sprintf('"%s" failed.', $test->name), "\n\n";
echo sprintf('Parsed: %s', print_r($output, TRUE)), "\n";
echo sprintf('Expected: %s', print_r($expected_output, TRUE)), "\n";
$this->tests_failed++;
}
# end if
}
# end if
}
# end loop
return TRUE;
}
/**
* @group parseP
*/
public function testParsePReturnsEmptyStringForBrHr()
{
$input = '<div class="h-card"><br class="p-name"/></div><div class="h-card"><hr class="p-name"/></div>';
$parser = new Parser($input);
$output = $parser->parse();
$this->assertArrayHasKey('name', $output['items'][0]['properties']);
$this->assertEquals('', $output['items'][0]['properties']['name'][0]);
$this->assertEquals('', $output['items'][0]['properties']['name'][0]);
}
/**
* @see https://github.com/indieweb/php-mf2/issues/27
*/
public function testParsesValueTitleDatetimes()
{
$input = <<<EOT
<div class="h-entry">
<h1 class="p-name">test</h1>
<span class="dt-published"><span class="value-title" title="2012-02-16T16:14:47+00:00"> </span>16.02.2012</span>
</div>
EOT;
$parser = new Parser($input);
$output = $parser->parse();
$this->assertEquals('2012-02-16T16:14:47+00:00', $output['items'][0]['properties']['published'][0]);
}
/**
* Wraps mfNamesFromClass to handle an element as input (common)
*
* @param DOMElement $e The element to get the classname for
* @param string $prefix The prefix to look for
* @return mixed See return value of mf2\Parser::mfNameFromClass()
*/
static function mfNamesFromElement(\DOMElement $e, $prefix = 'h-')
{
$class = $e->getAttribute('class');
return Parser::mfNamesFromClass($class, $prefix);
}
function detectBloggingSoftware($response)
{
$d = new MfParser($response->getBody(1), $response->getEffectiveUrl());
foreach ($d->query('//meta[@name="generator"]') as $generatorEl) {
if (stristr($generatorEl->getAttribute('content'), 'wordpress') !== false) {
return 'wordpress';
}
if (stristr($generatorEl->getAttribute('content'), 'mediawiki') !== false) {
return 'mediawiki';
}
if (stristr($generatorEl->getAttribute('content'), 'idno') !== false) {
return 'idno';
}
}
return null;
}
/**
* @group parseDT
* @group valueClass
*/
public function testYYYY_MM_DD__HHpm()
{
$input = '<div class="h-event"><span class="dt-start"><span class="value">2012-10-07</span> at <span class="value">9pm</span></span></div>';
$parser = new Parser($input);
$output = $parser->parse();
$this->assertArrayHasKey('start', $output['items'][0]['properties']);
$this->assertEquals('2012-10-07T21:00', $output['items'][0]['properties']['start'][0]);
}
public function testParsesNestedMicroformatsWithClassnamesInAnyOrder()
{
$input = <<<EOT
<div class="h-entry">
\t<div class="note- p-in-reply-to h-entry">Name</div>
</div>
EOT;
$parser = new Parser($input);
$output = $parser->parse();
$this->assertCount(1, $output['items'][0]['properties']['in-reply-to']);
$this->assertEquals('Name', $output['items'][0]['properties']['in-reply-to'][0]['properties']['name'][0]);
}
/**
* @group parseU
*/
public function testParseUHandlesSource()
{
$input = '<div class="h-entry"><video><source class="u-video" src="http://example.com/video.mp4" type="video/mp4"><source class="u-video" src="http://example.com/video.ogg" type="video/ogg"></video></div>';
$parser = new Parser($input);
$output = $parser->parse();
$this->assertArrayHasKey('video', $output['items'][0]['properties']);
$this->assertEquals('http://example.com/video.mp4', $output['items'][0]['properties']['video'][0]);
$this->assertEquals('http://example.com/video.ogg', $output['items'][0]['properties']['video'][1]);
}
/**
* @group baseUrl
*/
public function testResolvesRelativeBaseRelativeUrlsInImpliedMicroformats()
{
$input = '<base href="things/"/><a class="h-card"><img src="image.png" /></a>';
$parser = new Parser($input, 'http://example.com/');
$output = $parser->parse();
$this->assertEquals('http://example.com/things/image.png', $output['items'][0]['properties']['photo'][0]);
}
/**
* Kicks off the parsing routine
*
* If `$convertClassic` is set, any angle brackets in the results from non e-* properties
* will be HTML-encoded, bringing all output to the same level of encoding.
*
* If a DOMElement is set as the $context, only descendants of that element will
* be parsed for microformats.
*
* @param bool $convertClassic Whether or not to html-encode non e-* properties. Defaults to false
* @param \DOMElement $context Optional: An element from which to parse microformats
* @return array An array containing all the µfs found in the current document
*/
public function parse($convertClassic = true, \DOMElement $context = null)
{
$results = parent::parse($convertClassic, $context);
$results['items'] = $this->_refineResults($results['items']);
return $results;
}
/**
* @group parseU
*/
public function testParseUWithSpaces()
{
$input = '<div class="h-card"><a class="u-url" href=" http://example.com ">Awesome example website</a></div>';
$parser = new Parser($input);
$output = $parser->parse();
$this->assertArrayHasKey('url', $output['items'][0]['properties']);
$this->assertEquals('http://example.com', $output['items'][0]['properties']['url'][0]);
}
/**
* @group parseDT
* @group valueClass
*/
public function testDTStartDateOnly()
{
$input = '<div class="h-event"> <span class="dt-start"><span class="value">2014-06-07</span> </span> </div>';
$parser = new Parser($input);
$output = $parser->parse();
$this->assertArrayHasKey('start', $output['items'][0]['properties']);
$this->assertEquals('2014-06-07', $output['items'][0]['properties']['start'][0]);
}
/**
* @param $html
* @param $url
* @param $headers
* @param $topic
* @return array
*
* Currently $headers is just returned as-is, so is usually an array. This should probably have some abstraction for
* case insensitivity but that might be an issue if we can’t guarantee that $headers is an array, or that each item is a
* scalar or array — e.g. how are multiple values for a header handled?
*/
function contextFromResponse($html, $url, $headers, $topic)
{
$content = $html;
// TODO: per-host handling of microformats shim parsing should be handled by Mf2\Shim automatically.
$host = parse_url($url, PHP_URL_HOST);
if (in_array($host, ['www.twitter.com', 'twitter.com'])) {
$parser = new Mf2\Shim\Twitter($html, $url);
} elseif (in_array($host, ['www.facebook.com', 'facebook.com'])) {
$parser = new Mf2\Shim\Facebook($html, $url);
} else {
$parser = new Mf2\Parser($html, $url);
}
$mf2 = $parser->parse();
return ['topic' => $topic, 'url' => $url, 'mf2' => $mf2, 'content' => $content, 'domdocument' => $parser->doc, 'parser' => $parser, 'headers' => $headers];
}
public function testWhitespaceBetweenElements()
{
$input = <<<EOT
<div class="h-entry">
\t<data class="p-rsvp" value="yes">I'm attending</data>
\t<a class="u-in-reply-to" href="https://snarfed.org/2014-06-16_homebrew-website-club-at-quip">Homebrew Website Club at Quip</a>
\t<div class="p-content">Thanks for hosting!</div>
</div>
EOT;
$parser = new Parser($input);
$output = $parser->parse();
$this->assertContains('h-entry', $output['items'][0]['type']);
$this->assertNotContains('attendingHomebrew', $output['items'][0]['properties']['name'][0]);
}
public function testParseHcardInCategory()
{
$input = <<<EOT
<span class="h-entry">
\t<a class="p-author h-card" href="http://a.example.com/">Alice</a> tagged
\t<a href="http://b.example.com/" class="u-category h-card">Bob Smith</a> in
\t<a class="u-tag-of u-in-reply-to" href="http://s.example.com/permalink47">
\t\t<img src="http://s.example.com/photo47.png" alt="a photo of Bob and Cole" />
\t</a>
</span>
EOT;
$parser = new Parser($input);
$output = $parser->parse();
$this->assertContains('h-entry', $output['items'][0]['type']);
$this->assertArrayHasKey('category', $output['items'][0]['properties']);
$this->assertContains('h-card', $output['items'][0]['properties']['category'][0]['type']);
$this->assertArrayHasKey('name', $output['items'][0]['properties']['category'][0]['properties']);
$this->assertEquals('Bob Smith', $output['items'][0]['properties']['category'][0]['properties']['name'][0]);
$this->assertArrayHasKey('url', $output['items'][0]['properties']['category'][0]['properties']);
$this->assertEquals('http://b.example.com/', $output['items'][0]['properties']['category'][0]['properties']['url'][0]);
}
private function process($source, $target)
{
$this->_logger->log("Processing pingback from {$source} for {$target}", Zend_Log::INFO);
$curl = curl_init();
curl_setopt($curl, CURLOPT_URL, $source);
curl_setopt($curl, CURLOPT_HEADER, false);
curl_setopt($curl, CURLOPT_RETURNTRANSFER, true);
curl_setopt($curl, CURLOPT_USERAGENT, 'Storytlr/1.0');
curl_setopt($curl, CURLOPT_FOLLOWLOCATION, TRUE);
$response = curl_exec($curl);
$http_code = curl_getinfo($curl, CURLINFO_HTTP_CODE);
curl_close($curl);
if ($http_code != 200) {
$this->_logger->log("Failed to get content for {$source}", Zend_Log::DEBUG);
return;
}
// Parse the source page for microformats content
$parser = new Parser($response);
$output = $parser->parse();
$this->_logger->log("Parsed output: " . var_export($output, true), Zend_Log::DEBUG);
// Extract relevant data
$hcards = array();
$hentries = array();
$this->processItems($output["items"], $hcards, $hentries);
$this->_logger->log("Extracted hcards: " . var_export($hcards, true), Zend_Log::DEBUG);
$this->_logger->log("Extracted hentries: " . var_export($hentries, true), Zend_Log::DEBUG);
// Lookup if existing entry
preg_match('/(?P<source>\\d+)\\-(?P<item>\\d+)\\.html$/', $target, $matches);
$this->_logger->log("Matches: " . var_export($matches, true), Zend_Log::DEBUG);
$source_id = $matches["source"];
$item_id = $matches["item"];
// Get the source and the user owning it
$data = new Data();
$sources = new Sources();
$users = new Users();
// Does it relate to an item ?
if ($source_id && $item_id) {
$s = $sources->getSource($source_id);
$i = $data->getItem($source_id, $item_id);
if ($s && $i) {
$user = $users->getUser($s['user_id']);
}
}
// Otherwise, can we relate to a user ?
if (!$user) {
$user = $this->lookupUser($target);
}
// No user ? We have to giveup
if (!$user) {
throw new Exception('Failed to find corresponding storytlr user.');
}
// Is this a h-entry ?
if (count($hentries) > 0) {
$hentry = $hentries[0];
if (count($hentry["author"]) > 0) {
$hcard = $hentry["author"][0];
}
}
// If no hcard yet (maybe there was no h-entry, we grab the top-level one
if (!$hcard && count($hcards) > 0) {
$hcard = $hcards[0];
}
// Find the published date
if ($hentry && $hentry["published"]) {
$timestamp = strtotime($hentry["published"]);
}
// If no timestamp, we fallback to now
if (!$timestamp) {
$timestamp = time();
}
// Add the mention to the database
try {
$mentions = new Mentions();
$mentions->addMention($source_id, $item_id, $user->id, $source, $hentry["title"], $hcard["name"], $hcard["url"], "", $hcard["avatar"], $timestamp);
} catch (Exception $e) {
$this->_logger->log("Exception when storing mention: " . $e->getMessage(), Zend_Log::ERR);
return;
}
// Send an email alert to owner
try {
$on_comment = $this->_properties->getProperty('on_comment');
if ($on_comment) {
Stuffpress_Emails::sendMentionEmail($user->email, $user->username, $hcard["name"], $hcard["url"], $hentry["title"], $source, $target);
}
} catch (Exception $e) {
$this->_logger->log("Sending mention notification exception: " . $e->getMessage(), Zend_Log::ERR);
}
}
/**
* From http://microformats.org/wiki/microformats-2
*/
public function testMoreDetailedPerson()
{
$input = '<div class="h-card">
<img class="u-photo" alt="photo of Mitchell"
src="https://webfwd.org/content/about-experts/300.mitchellbaker/mentor_mbaker.jpg"/>
<a class="p-name u-url"
href="http://blog.lizardwrangler.com/"
>Mitchell Baker</a>
(<a class="u-url"
href="https://twitter.com/MitchellBaker"
>@MitchellBaker</a>)
<span class="p-org">Mozilla Foundation</span>
<p class="p-note">
Mitchell is responsible for setting the direction and scope of the Mozilla Foundation and its activities.
</p>
<span class="p-category">Strategy</span>
<span class="p-category">Leadership</span>
</div>';
$expected = '{
"rels": {},
"items": [{
"type": ["h-card"],
"properties": {
"photo": ["https://webfwd.org/content/about-experts/300.mitchellbaker/mentor_mbaker.jpg"],
"name": ["Mitchell Baker"],
"url": [
"http://blog.lizardwrangler.com/",
"https://twitter.com/MitchellBaker"
],
"org": ["Mozilla Foundation"],
"note": ["Mitchell is responsible for setting the direction and scope of the Mozilla Foundation and its activities."],
"category": [
"Strategy",
"Leadership"
]
}
}]
}';
$parser = new Parser($input);
$output = $parser->parse();
$this->assertJsonStringEqualsJsonString(json_encode($output), $expected);
}
public function testParsesFBerrimanClassicHEntry()
{
$input = <<<EOT
<article id="post-976" class="post-976 post type-post status-publish format-standard hentry category-speaking category-web-dev tag-conferences tag-front-trends tag-fronttrends tag-speaking tag-txjs">
\t<header class="entry-header">
\t\t<h1 class="entry-title">
\t\t\t<a href="http://fberriman.com/2013/05/14/april-recap-txjs-front-trends/" rel="bookmark">April recap – TXJS & Front-Trends</a>
\t\t</h1>
\t\t
\t\t<div class="entry-meta">
\t\t\t<span class="date">
\t\t\t\t<a href="http://fberriman.com/2013/05/14/april-recap-txjs-front-trends/" title="Permalink to April recap – TXJS & Front-Trends" rel="bookmark">
\t\t\t\t\t<time class="entry-date" datetime="2013-05-14T11:54:06+00:00">May 14, 2013</time>
\t\t\t\t</a>
\t\t\t</span>
\t\t\t<span class="categories-links">
\t\t\t\t<a href="http://fberriman.com/category/speaking/" title="View all posts in Speaking" rel="category tag">Speaking</a>,
\t\t\t\t<a href="http://fberriman.com/category/web-dev/" title="View all posts in Web Dev" rel="category tag">Web Dev</a>
\t\t\t</span>
\t\t\t<span class="tags-links">
\t\t\t\t<a href="http://fberriman.com/tag/conferences/" rel="tag">conferences</a>,
\t\t\t\t<a href="http://fberriman.com/tag/front-trends/" rel="tag">front-trends</a>,
\t\t\t\t<a href="http://fberriman.com/tag/fronttrends/" rel="tag">fronttrends</a>,
\t\t\t\t<a href="http://fberriman.com/tag/speaking/" rel="tag">Speaking</a>,
\t\t\t\t<a href="http://fberriman.com/tag/txjs/" rel="tag">txjs</a>
\t\t\t</span>
\t\t\t<span class="author vcard"><a class="url fn n" href="http://fberriman.com/author/admin/" title="View all posts by Frances" rel="author">Frances</a></span>\t\t\t\t\t</div>
\t</header>
\t\t<div class="entry-content">
\t\t<p>April was pretty decent. I got to attend two very good conferences <strong>and</strong> I got to speak at them.</p>
\t\t\t</div>
\t
\t<footer class="entry-meta">
\t\t<div class="comments-link">
\t\t\t<a href="http://fberriman.com/2013/05/14/april-recap-txjs-front-trends/#respond" title="Comment on April recap – TXJS & Front-Trends"><span class="leave-reply">Leave a comment</span></a>
\t\t</div>
\t</footer><!-- .entry-meta -->
</article><!-- #post -->
EOT;
$parser = new Parser($input);
$result = $parser->parse();
$e = $result['items'][0];
$this->assertContains('h-entry', $e['type']);
}
/**
* Parse Microformats2
*
* Functional shortcut for the commonest cases of parsing microformats2 from HTML.
*
* Example usage:
*
* use Mf2;
* $output = Mf2\parse('<span class="h-card">Barnaby Walters</span>');
* echo json_encode($output, JSON_PRETTY_PRINT);
*
* Produces:
*
* {
* "items": [
* {
* "type": ["h-card"],
* "properties": {
* "name": ["Barnaby Walters"]
* }
* }
* ],
* "rels": {}
* }
*
* @param string|DOMDocument $input The HTML string or DOMDocument object to parse
* @param string $url The URL the input document was found at, for relative URL resolution
* @param bool $convertClassic whether or not to convert classic microformats
* @return array Canonical MF2 array structure
*/
function parse($input, $url = null, $convertClassic = true)
{
$parser = new Parser($input, $url);
return $parser->parse($convertClassic);
}
/**
* rel-me links
* given the HTML and URL of a page, returns all the rel-me links found on that page
* @return array
*/
function relMeLinks($html, $url)
{
$parser = new Mf2\Parser($html, $url);
$mf = $parser->parse();
$relMeLinks = @($mf['rels']['me'] ?: array());
return array_unique($relMeLinks);
}
/**
* From http://microformats.org/wiki/microformats2#combining_microformats
*/
public function testHCardChildHCard()
{
$input = '<div class="h-card">
<a class="p-name u-url"
href="http://blog.lizardwrangler.com/">
Mitchell Baker</a>
(<a class="h-card h-org" href="http://mozilla.org/">
Mozilla Foundation</a>)
</div>';
$expected = '{
"rels": {},
"items": [{
"type": ["h-card"],
"properties": {
"name": ["Mitchell Baker"],
"url": ["http://blog.lizardwrangler.com/"]
},
"children": [{
"type": ["h-card","h-org"],
"properties": {
"name": ["Mozilla Foundation"],
"url": ["http://mozilla.org/"]
}
}]
}]
}';
$parser = new Parser($input);
$output = $parser->parse();
$this->assertJsonStringEqualsJsonString(json_encode($output), $expected);
}
/**
* generate the comment data from the microformatted content
*
* @param WP_Comment $commentdata the comment object
* @param string $target the target url
* @param string $html the parsed html
*
* @return array
*/
public static function generate_commentdata($commentdata, $target, $html)
{
global $wpdb;
// add source
$source = $commentdata['comment_author_url'];
// parse source html
$parser = new Parser($html, $source);
$mf_array = $parser->parse(true);
// get all 'relevant' entries
$entries = self::get_entries($mf_array);
if (empty($entries)) {
return array();
}
// get the entry of interest
$entry = self::get_representative_entry($entries, $target);
if (empty($entry)) {
return array();
}
// the entry properties
$properties = $entry['properties'];
// try to find some content
// @link http://indiewebcamp.com/comments-presentation
if (self::check_mf_attr('summary', $properties)) {
$commentdata['comment_content'] = wp_slash($properties['summary'][0]);
} elseif (self::check_mf_attr('content', $properties)) {
$commentdata['comment_content'] = wp_filter_kses($properties['content'][0]['html']);
} elseif (self::check_mf_attr('name', $properties)) {
$commentdata['comment_content'] = wp_slash($properties['name'][0]);
}
$commentdata['comment_content'] = trim($commentdata['comment_content']);
// set the right date
if (self::check_mf_attr('published', $properties)) {
$time = strtotime($properties['published'][0]);
$commentdata['comment_date'] = get_date_from_gmt(date('Y-m-d H:i:s', $time), 'Y-m-d H:i:s');
} elseif (self::check_mf_attr('updated', $properties)) {
$time = strtotime($properties['updated'][0]);
$commentdata['comment_date'] = get_date_from_gmt(date('Y-m-d H:i:s', $time), 'Y-m-d H:i:s');
}
$author = null;
// check if h-card has an author
if (isset($properties['author']) && isset($properties['author'][0]['properties'])) {
$author = $properties['author'][0]['properties'];
} else {
$author = self::get_representative_author($mf_array, $source);
}
// if author is present use the informations for the comment
if ($author) {
if (self::check_mf_attr('name', $author)) {
$commentdata['comment_author'] = wp_slash($author['name'][0]);
}
if (self::check_mf_attr('email', $author)) {
$commentdata['comment_author_email'] = wp_slash($author['email'][0]);
}
if (self::check_mf_attr('url', $author)) {
$commentdata['_author_url'] = esc_url_raw($author['url'][0]);
}
if (self::check_mf_attr('photo', $author)) {
$commentdata['_avatar'] = esc_url_raw($author['photo'][0]);
}
}
// set canonical url (u-url)
if (self::check_mf_attr('url', $properties)) {
$commentdata['_canonical'] = esc_url_raw($properties['url'][0]);
} else {
$commentdata['_canonical'] = esc_url_raw($source);
}
// check rsvp property
if (self::check_mf_attr('rsvp', $properties)) {
$commentdata['_type'] = wp_slash('rsvp:' . $properties['rsvp'][0]);
} else {
// get post type
$commentdata['_type'] = wp_slash(self::get_entry_type($target, $entry, $mf_array));
}
return $commentdata;
}
public function testMultipleImpliedHCards()
{
$input = '<span class="h-card">Frances Berriman</span>
<a class="h-card" href="http://benward.me">Ben Ward</a>
<img class="h-card" alt="Sally Ride"
src="http://upload.wikimedia.org/wikipedia/commons/a/a4/Ride-s.jpg"/>
<a class="h-card" href="http://tantek.com">
<img alt="Tantek Çelik" src="http://ttk.me/logo.jpg"/>
</a>';
$expected = '{
"rels": {},
"items": [{
"type": ["h-card"],
"properties": {
"name": ["Frances Berriman"]
}
},
{
"type": ["h-card"],
"properties": {
"name": ["Ben Ward"],
"url": ["http://benward.me"]
}
},
{
"type": ["h-card"],
"properties": {
"name": ["Sally Ride"],
"photo": ["http://upload.wikimedia.org/wikipedia/commons/a/a4/Ride-s.jpg"]
}
},
{
"type": ["h-card"],
"properties": {
"name": ["Tantek Çelik"],
"url": ["http://tantek.com"],
"photo": ["http://ttk.me/logo.jpg"]
}
}]
}';
$parser = new Parser($input, '', true);
$output = $parser->parse();
$this->assertJsonStringEqualsJsonString(json_encode($output), $expected);
}
<?php
namespace IndieWeb;
use mf2\Parser as Mf2Parser;
use BarnabyWalters\Mf2;
ob_start();
require __DIR__ . '/bootstrap.php';
ob_end_clean();
if (PHP_SAPI === 'cli') {
$html = file_get_contents(__DIR__ . '/rel-me-test.html');
$parser = new Mf2Parser($html);
$mf = $parser->parse();
$testSuites = Mf2\findMicroformatsByType($mf, 'h-x-test-suite');
if (count($testSuites) === 0) {
die("Found no test suites in rel-me-test.html");
}
$testSuite = $testSuites[0];
echo "\n" . Mf2\getProp($testSuite, 'name') . "\n";
echo "============================\n";
foreach ($testSuite['properties']['x-test-case'] as $testCase) {
echo "\n";
$params = $testCase['properties']['x-parameter'];
$meUrl = array_shift($params);
if (count($params) === 0) {
$redirects = mockFollowOneRedirect(array(null));
} else {
$redirects = mockFollowOneRedirect($params);
}
list($expectedUrl, $expectedSecure) = $testCase['properties']['x-expected-result'];
$expectedSecure = $expectedSecure === 'true';