diff options
author | Andrew Dolgov <[email protected]> | 2017-05-24 23:43:43 +0300 |
---|---|---|
committer | Andrew Dolgov <[email protected]> | 2017-05-24 23:43:43 +0300 |
commit | f025e8e46e34221efc19e9b5ed1a53f7f0ec2b49 (patch) | |
tree | daa4e8b2138627f163b9f79622b3ceed9f2ecec6 | |
parent | 9e381bc2021dc427f62c463dac9cf7a82a66616f (diff) |
test implementation for the most unnecessary thing ever: jsonfeed
-rw-r--r-- | classes/feeditem/json.php | 88 | ||||
-rw-r--r-- | classes/feedparser.php | 229 | ||||
-rw-r--r-- | classes/rssutils.php | 2 |
3 files changed, 216 insertions, 103 deletions
diff --git a/classes/feeditem/json.php b/classes/feeditem/json.php new file mode 100644 index 000000000..c9ecd59e4 --- /dev/null +++ b/classes/feeditem/json.php @@ -0,0 +1,88 @@ +<?php +class FeedItem_Json extends FeedItem_Common { + + /* for JSON feed only $elem is passed which is the actual entry as an object */ + function __construct($elem, $doc, $xpath) { + $this->elem = $elem; + $this->doc = $doc; + } + + function get_id() { + return $this->elem->id; + } + + function get_date() { + return isset($this->elem->date_published) ? strtotime($this->elem->date_published) : false; + } + + function get_author() { + $author = false; + + if (isset($this->author)) { + $author = $this->author; + } else if (isset($this->doc->author)) { + $author = $this->doc->author; + } + + if ($author && $author->name) { + return $author->name; + } + } + + function get_comments_url() { + return false; + } + + function get_comments_count() { + return false; + } + + function get_link() { + return isset($this->elem->url) ? $this->elem->url : false; + } + + function get_title() { + return $this->elem->title; + } + + function get_content() { + return isset($this->elem->content_html) ? $this->elem->content_html : $this->elem->content_text; + } + + function get_description() { + return isset($this->elem->summary) ? $this->elem->summary : false; + } + + function get_categories() { + $cats = array(); + + if (is_array($this->elem->tags)) { + foreach ($this->elem->tags as $cat) { + array_push($cats, trim($cat)); + } + } + + return $cats; + } + + function get_enclosures() { + $encs = array(); + + if (is_array($this->elem->attachments)) { + foreach ($this->elem->attachments as $enclosure) { + + $enc = new FeedEnclosure(); + + $enc->type = $enclosure["mime_type"]; + $enc->link = $enclosure["url"]; + @$enc->length = $enclosure["size_in_bytes"]; + + array_push($encs, $enc); + + } + } + + return $encs; + } + +}
\ No newline at end of file diff --git a/classes/feedparser.php b/classes/feedparser.php index 860ebd73f..b416808be 100644 --- a/classes/feedparser.php +++ b/classes/feedparser.php @@ -3,15 +3,17 @@ class FeedParser { private $doc; private $error; private $libxml_errors = array(); - private $items; + private $items = array(); private $link; private $title; private $type; private $xpath; + private $jsonobj; const FEED_RDF = 0; const FEED_RSS = 1; const FEED_ATOM = 2; + const FEED_JSON = 3; function normalize_encoding($data) { if (preg_match('/^(<\?xml[\t\n\r ].*?encoding[\t\n\r ]*=[\t\n\r ]*["\'])(.+?)(["\'].*?\?>)/s', $data, $matches) === 1) { @@ -28,6 +30,17 @@ class FeedParser { } function __construct($data) { + $this->jsonobj = @json_decode($data, false); + + // not sure of a better solution to report parsing errors for both kinds of (potentially broken) content + + if ($this->jsonobj) { + return; + } else if (!preg_match("/<(\\?xml|feed|channel|rdf|rss)/m", mb_substr($data, 0, 512))) { + $this->error = 'JSON error: ' . json_last_error_msg(); + return; + } + libxml_use_internal_errors(true); libxml_clear_errors(); $this->doc = new DOMDocument(); @@ -87,144 +100,156 @@ class FeedParser { } } libxml_clear_errors(); - - $this->items = array(); } function init() { - $root = $this->doc->firstChild; - $xpath = new DOMXPath($this->doc); - $xpath->registerNamespace('atom', 'http://www.w3.org/2005/Atom'); - $xpath->registerNamespace('atom03', 'http://purl.org/atom/ns#'); - $xpath->registerNamespace('media', 'http://search.yahoo.com/mrss/'); - $xpath->registerNamespace('rdf', 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'); - $xpath->registerNamespace('slash', 'http://purl.org/rss/1.0/modules/slash/'); - $xpath->registerNamespace('dc', 'http://purl.org/dc/elements/1.1/'); - $xpath->registerNamespace('content', 'http://purl.org/rss/1.0/modules/content/'); - $xpath->registerNamespace('thread', 'http://purl.org/syndication/thread/1.0'); - - $this->xpath = $xpath; - - $root = $xpath->query("(//atom03:feed|//atom:feed|//channel|//rdf:rdf|//rdf:RDF)"); - - if ($root && $root->length > 0) { - $root = $root->item(0); - - if ($root) { - switch (mb_strtolower($root->tagName)) { - case "rdf:rdf": - $this->type = $this::FEED_RDF; - break; - case "channel": - $this->type = $this::FEED_RSS; - break; - case "feed": - case "atom:feed": - $this->type = $this::FEED_ATOM; - break; - default: - if( !isset($this->error) ){ - $this->error = "Unknown/unsupported feed type"; + + if ($this->jsonobj) { + $this->type = $this::FEED_JSON; + + $this->title = $this->jsonobj->title; + $this->link = $this->jsonobj->feed_url; + + foreach ($this->jsonobj->items as $item) { + array_push($this->items, new FeedItem_Json($item, $this->jsonobj, false)); + } + + } else if ($this->doc) { + + $xpath = new DOMXPath($this->doc); + $xpath->registerNamespace('atom', 'http://www.w3.org/2005/Atom'); + $xpath->registerNamespace('atom03', 'http://purl.org/atom/ns#'); + $xpath->registerNamespace('media', 'http://search.yahoo.com/mrss/'); + $xpath->registerNamespace('rdf', 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'); + $xpath->registerNamespace('slash', 'http://purl.org/rss/1.0/modules/slash/'); + $xpath->registerNamespace('dc', 'http://purl.org/dc/elements/1.1/'); + $xpath->registerNamespace('content', 'http://purl.org/rss/1.0/modules/content/'); + $xpath->registerNamespace('thread', 'http://purl.org/syndication/thread/1.0'); + + $this->xpath = $xpath; + + $root = $xpath->query("(//atom03:feed|//atom:feed|//channel|//rdf:rdf|//rdf:RDF)"); + + if ($root && $root->length > 0) { + $root = $root->item(0); + + if ($root) { + switch (mb_strtolower($root->tagName)) { + case "rdf:rdf": + $this->type = $this::FEED_RDF; + break; + case "channel": + $this->type = $this::FEED_RSS; + break; + case "feed": + case "atom:feed": + $this->type = $this::FEED_ATOM; + break; + default: + if( !isset($this->error) ){ + $this->error = "Unknown/unsupported feed type"; + } + return; } - return; } - } - switch ($this->type) { - case $this::FEED_ATOM: + switch ($this->type) { + case $this::FEED_ATOM: - $title = $xpath->query("//atom:feed/atom:title")->item(0); + $title = $xpath->query("//atom:feed/atom:title")->item(0); - if (!$title) - $title = $xpath->query("//atom03:feed/atom03:title")->item(0); + if (!$title) + $title = $xpath->query("//atom03:feed/atom03:title")->item(0); - if ($title) { - $this->title = $title->nodeValue; - } + if ($title) { + $this->title = $title->nodeValue; + } - $link = $xpath->query("//atom:feed/atom:link[not(@rel)]")->item(0); + $link = $xpath->query("//atom:feed/atom:link[not(@rel)]")->item(0); - if (!$link) - $link = $xpath->query("//atom:feed/atom:link[@rel='alternate']")->item(0); + if (!$link) + $link = $xpath->query("//atom:feed/atom:link[@rel='alternate']")->item(0); - if (!$link) - $link = $xpath->query("//atom03:feed/atom03:link[not(@rel)]")->item(0); + if (!$link) + $link = $xpath->query("//atom03:feed/atom03:link[not(@rel)]")->item(0); - if (!$link) - $link = $xpath->query("//atom03:feed/atom03:link[@rel='alternate']")->item(0); + if (!$link) + $link = $xpath->query("//atom03:feed/atom03:link[@rel='alternate']")->item(0); - if ($link && $link->hasAttributes()) { - $this->link = $link->getAttribute("href"); - } + if ($link && $link->hasAttributes()) { + $this->link = $link->getAttribute("href"); + } - $articles = $xpath->query("//atom:entry"); + $articles = $xpath->query("//atom:entry"); - if (!$articles || $articles->length == 0) - $articles = $xpath->query("//atom03:entry"); + if (!$articles || $articles->length == 0) + $articles = $xpath->query("//atom03:entry"); - foreach ($articles as $article) { - array_push($this->items, new FeedItem_Atom($article, $this->doc, $this->xpath)); - } + foreach ($articles as $article) { + array_push($this->items, new FeedItem_Atom($article, $this->doc, $this->xpath)); + } - break; - case $this::FEED_RSS: - $title = $xpath->query("//channel/title")->item(0); + break; + case $this::FEED_RSS: + $title = $xpath->query("//channel/title")->item(0); - if ($title) { - $this->title = $title->nodeValue; - } + if ($title) { + $this->title = $title->nodeValue; + } - $link = $xpath->query("//channel/link")->item(0); + $link = $xpath->query("//channel/link")->item(0); - if ($link) { - if ($link->getAttribute("href")) - $this->link = $link->getAttribute("href"); - else if ($link->nodeValue) - $this->link = $link->nodeValue; - } + if ($link) { + if ($link->getAttribute("href")) + $this->link = $link->getAttribute("href"); + else if ($link->nodeValue) + $this->link = $link->nodeValue; + } - $articles = $xpath->query("//channel/item"); + $articles = $xpath->query("//channel/item"); - foreach ($articles as $article) { - array_push($this->items, new FeedItem_RSS($article, $this->doc, $this->xpath)); - } + foreach ($articles as $article) { + array_push($this->items, new FeedItem_RSS($article, $this->doc, $this->xpath)); + } - break; - case $this::FEED_RDF: - $xpath->registerNamespace('rssfake', 'http://purl.org/rss/1.0/'); + break; + case $this::FEED_RDF: + $xpath->registerNamespace('rssfake', 'http://purl.org/rss/1.0/'); - $title = $xpath->query("//rssfake:channel/rssfake:title")->item(0); + $title = $xpath->query("//rssfake:channel/rssfake:title")->item(0); - if ($title) { - $this->title = $title->nodeValue; - } + if ($title) { + $this->title = $title->nodeValue; + } - $link = $xpath->query("//rssfake:channel/rssfake:link")->item(0); + $link = $xpath->query("//rssfake:channel/rssfake:link")->item(0); - if ($link) { - $this->link = $link->nodeValue; - } + if ($link) { + $this->link = $link->nodeValue; + } - $articles = $xpath->query("//rssfake:item"); + $articles = $xpath->query("//rssfake:item"); - foreach ($articles as $article) { - array_push($this->items, new FeedItem_RSS($article, $this->doc, $this->xpath)); - } + foreach ($articles as $article) { + array_push($this->items, new FeedItem_RSS($article, $this->doc, $this->xpath)); + } - break; + break; - } + } - if ($this->title) $this->title = trim($this->title); - if ($this->link) $this->link = trim($this->link); + if ($this->title) $this->title = trim($this->title); + if ($this->link) $this->link = trim($this->link); - } else { - if( !isset($this->error) ){ - $this->error = "Unknown/unsupported feed type"; + } else { + if( !isset($this->error) ){ + $this->error = "Unknown/unsupported feed type"; + } + return; } - return; } + } function format_error($error) { diff --git a/classes/rssutils.php b/classes/rssutils.php index e6fe39592..6d49d6383 100644 --- a/classes/rssutils.php +++ b/classes/rssutils.php @@ -339,7 +339,7 @@ class RSSUtils { $date_feed_processed = date('Y-m-d H:i'); - $cache_filename = CACHE_DIR . "/simplepie/" . sha1($fetch_url) . ".xml"; + $cache_filename = CACHE_DIR . "/simplepie/" . sha1($fetch_url); $pluginhost = new PluginHost(); $pluginhost->set_debug($debug_enabled); |