]>
git.wh0rd.org - tt-rss.git/blob - classes/feedparser.php
5 private $libxml_errors = array();
16 function normalize_encoding($data) {
17 if (preg_match('/^(<\?xml[\t\n\r ].*?encoding[\t\n\r ]*=[\t\n\r ]*["\'])(.+?)(["\'].*?\?>)/s', $data, $matches) === 1) {
18 $data = mb_convert_encoding($data, 'UTF-8', $matches[2]);
20 $data = preg_replace('/^<\?xml[\t\n\r ].*?\?>/s', $matches[1] . "UTF-8" . $matches[3] , $data);
26 function __construct($data) {
27 libxml_use_internal_errors(true);
28 libxml_clear_errors();
29 $this->doc
= new DOMDocument();
30 $this->doc
->loadXML($data);
32 mb_substitute_character("none");
34 $error = libxml_get_last_error();
36 // libxml compiled without iconv?
37 if ($error && $error->code
== 32) {
38 $data = $this->normalize_encoding($data);
41 libxml_clear_errors();
43 $this->doc
= new DOMDocument();
44 $this->doc
->loadXML($data);
46 $error = libxml_get_last_error();
50 // some terrible invalid unicode entity?
52 foreach (libxml_get_errors() as $err) {
53 if ($err->code
== 9) {
54 // if the source feed is not in utf8, next conversion will fail
55 $data = $this->normalize_encoding($data);
57 // remove dangling bytes
58 $data = mb_convert_encoding($data, 'UTF-8', 'UTF-8');
60 // apparently not all UTF-8 characters are valid for XML
61 $data = preg_replace('/[^\x{0009}\x{000a}\x{000d}\x{0020}-\x{D7FF}\x{E000}-\x{FFFD}]+/u', ' ', $data);
64 libxml_clear_errors();
66 $this->doc
= new DOMDocument();
67 $this->doc
->loadXML($data);
69 $error = libxml_get_last_error();
77 foreach (libxml_get_errors() as $error) {
78 if ($error->level
== LIBXML_ERR_FATAL
) {
79 if(!isset($this->error
)) //currently only the first error is reported
80 $this->error
= $this->format_error($error);
81 $this->libxml_errors
[] = $this->format_error($error);
85 libxml_clear_errors();
87 $this->items
= array();
91 $root = $this->doc
->firstChild
;
92 $xpath = new DOMXPath($this->doc
);
93 $xpath->registerNamespace('atom', 'http://www.w3.org/2005/Atom');
94 $xpath->registerNamespace('atom03', 'http://purl.org/atom/ns#');
95 $xpath->registerNamespace('media', 'http://search.yahoo.com/mrss/');
96 $xpath->registerNamespace('rdf', 'http://www.w3.org/1999/02/22-rdf-syntax-ns#');
97 $xpath->registerNamespace('slash', 'http://purl.org/rss/1.0/modules/slash/');
98 $xpath->registerNamespace('dc', 'http://purl.org/dc/elements/1.1/');
99 $xpath->registerNamespace('content', 'http://purl.org/rss/1.0/modules/content/');
100 $xpath->registerNamespace('thread', 'http://purl.org/syndication/thread/1.0');
102 $this->xpath
= $xpath;
104 $root = $xpath->query("(//atom03:feed|//atom:feed|//channel|//rdf:rdf|//rdf:RDF)");
106 if ($root && $root->length
> 0) {
107 $root = $root->item(0);
110 switch (mb_strtolower($root->tagName
)) {
112 $this->type
= $this::FEED_RDF
;
115 $this->type
= $this::FEED_RSS
;
119 $this->type
= $this::FEED_ATOM
;
122 if( !isset($this->error
) ){
123 $this->error
= "Unknown/unsupported feed type";
129 switch ($this->type
) {
130 case $this::FEED_ATOM
:
132 $title = $xpath->query("//atom:feed/atom:title")->item(0);
135 $title = $xpath->query("//atom03:feed/atom03:title")->item(0);
139 $this->title
= $title->nodeValue
;
142 $link = $xpath->query("//atom:feed/atom:link[not(@rel)]")->item(0);
145 $link = $xpath->query("//atom:feed/atom:link[@rel='alternate']")->item(0);
148 $link = $xpath->query("//atom03:feed/atom03:link[not(@rel)]")->item(0);
151 $link = $xpath->query("//atom03:feed/atom03:link[@rel='alternate']")->item(0);
153 if ($link && $link->hasAttributes()) {
154 $this->link
= $link->getAttribute("href");
157 $articles = $xpath->query("//atom:entry");
159 if (!$articles ||
$articles->length
== 0)
160 $articles = $xpath->query("//atom03:entry");
162 foreach ($articles as $article) {
163 array_push($this->items
, new FeedItem_Atom($article, $this->doc
, $this->xpath
));
167 case $this::FEED_RSS
:
168 $title = $xpath->query("//channel/title")->item(0);
171 $this->title
= $title->nodeValue
;
174 $link = $xpath->query("//channel/link")->item(0);
177 if ($link->getAttribute("href"))
178 $this->link
= $link->getAttribute("href");
179 else if ($link->nodeValue
)
180 $this->link
= $link->nodeValue
;
183 $articles = $xpath->query("//channel/item");
185 foreach ($articles as $article) {
186 array_push($this->items
, new FeedItem_RSS($article, $this->doc
, $this->xpath
));
190 case $this::FEED_RDF
:
191 $xpath->registerNamespace('rssfake', 'http://purl.org/rss/1.0/');
193 $title = $xpath->query("//rssfake:channel/rssfake:title")->item(0);
196 $this->title
= $title->nodeValue
;
199 $link = $xpath->query("//rssfake:channel/rssfake:link")->item(0);
202 $this->link
= $link->nodeValue
;
205 $articles = $xpath->query("//rssfake:item");
207 foreach ($articles as $article) {
208 array_push($this->items
, new FeedItem_RSS($article, $this->doc
, $this->xpath
));
215 if ($this->title
) $this->title
= trim($this->title
);
216 if ($this->link
) $this->link
= trim($this->link
);
219 if( !isset($this->error
) ){
220 $this->error
= "Unknown/unsupported feed type";
226 function format_error($error) {
228 return sprintf("LibXML error %s at line %d (column %d): %s",
229 $error->code
, $error->line
, $error->column
,
241 return $this->libxml_errors
;
244 function get_link() {
248 function get_title() {
252 function get_items() {
256 function get_links($rel) {
259 switch ($this->type
) {
260 case $this::FEED_ATOM
:
261 $links = $this->xpath
->query("//atom:feed/atom:link");
263 foreach ($links as $link) {
264 if (!$rel ||
$link->hasAttribute('rel') && $link->getAttribute('rel') == $rel) {
265 array_push($rv, trim($link->getAttribute('href')));
269 case $this::FEED_RSS
:
270 $links = $this->xpath
->query("//atom:link");
272 foreach ($links as $link) {
273 if (!$rel ||
$link->hasAttribute('rel') && $link->getAttribute('rel') == $rel) {
274 array_push($rv, trim($link->getAttribute('href')));