1 | <?php |
---|
2 | /**************************************************************************\ |
---|
3 | * eGroupWare API - rss parser * |
---|
4 | * ------------------------------------------------------------------------ * |
---|
5 | * This is not part of eGroupWare, but is used by eGroupWare. * |
---|
6 | * http://www.egroupware.org/ * |
---|
7 | * ------------------------------------------------------------------------ * |
---|
8 | * This program is free software; you can redistribute it and/or modify it * |
---|
9 | * under the terms of the GNU Lesser General Public License as published * |
---|
10 | * by the Free Software Foundation; either version 2.1 of the License, or * |
---|
11 | * any later version. * |
---|
12 | \**************************************************************************/ |
---|
13 | |
---|
14 | |
---|
15 | /* |
---|
16 | * rssparse.php3 |
---|
17 | * |
---|
18 | * Copyright (C) 2000 Jeremey Barrett |
---|
19 | * j@nwow.org |
---|
20 | * http://nwow.org |
---|
21 | * |
---|
22 | * Version 0.4 |
---|
23 | * |
---|
24 | * This library is free software; you can redistribute it and/or modify |
---|
25 | * it under the terms of the GNU Lesser General Public License as published by |
---|
26 | * the Free Software Foundation; either version 2.1 of the License, or |
---|
27 | * (at your option) any later version. |
---|
28 | * |
---|
29 | * This library is distributed in the hope that it will be useful, |
---|
30 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
---|
31 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
---|
32 | * GNU Lesser General Public License for more details. |
---|
33 | * |
---|
34 | * You should have received a copy of the GNU Lesser General Public License |
---|
35 | * along with this program; if not, write to the Free Software |
---|
36 | * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
---|
37 | * |
---|
38 | * |
---|
39 | * rssparse.php3 is a small PHP script for parsing RDF/RSS XML data. It has been |
---|
40 | * tested with a number of popular web news and information sites such as |
---|
41 | * slashdot.org, lwn.net, and freshmeat.net. This is not meant to be exhaustive |
---|
42 | * but merely to provide the basic necessities. It will grow in capabilities |
---|
43 | * with time, I'm sure. |
---|
44 | * |
---|
45 | * This is code I wrote for Nerds WithOut Wires, http://nwow.org. |
---|
46 | * |
---|
47 | * |
---|
48 | * USAGE: |
---|
49 | * In your PHP script, simply do something akin to this: |
---|
50 | * |
---|
51 | * include("rssparse.php3"); |
---|
52 | * $fp = fopen("file", "r"); |
---|
53 | * $rss = rssparse($fp); |
---|
54 | * |
---|
55 | * if (!$rss) { |
---|
56 | * ERROR; |
---|
57 | * } |
---|
58 | * |
---|
59 | * while (list(,$item) = each($rss->items)) { |
---|
60 | * printf("Title: %s\n", $item["title"]); |
---|
61 | * printf("Link: %s\n", $item["link"]); |
---|
62 | * printf("Description: %s\n", $item["desc"]); |
---|
63 | * } |
---|
64 | * |
---|
65 | * printf("Channel Title: %s\n", $rss->title); |
---|
66 | * printf("Channel Description: %s\n", $rss->desc); |
---|
67 | * printf("Channel Link: %s\n", $rss->link); |
---|
68 | * |
---|
69 | * printf("Image Title: %s\n", $rss->image["title"]); |
---|
70 | * printf("Image URL: %s\n", $rss->image["url"]); |
---|
71 | * printf("Image Description: %s\n", $rss->image["desc"]); |
---|
72 | * printf("Image Link: %s\n", $rss->image["link"]); |
---|
73 | * |
---|
74 | * |
---|
75 | * CHANGES: |
---|
76 | * 0.4 - rssparse.php3 now supports the channel image tag and correctly supports |
---|
77 | * RSS-style channels. |
---|
78 | * |
---|
79 | * |
---|
80 | * BUGS: |
---|
81 | * Width and height tags in image not supported, some other tags not supported |
---|
82 | * yet. |
---|
83 | * |
---|
84 | * |
---|
85 | * IMPORTANT NOTE: |
---|
86 | * This requires PHP's XML routines. You must configure PHP with --with-xml. |
---|
87 | */ |
---|
88 | |
---|
89 | function _rssparse_start_elem ($parser, $elem, $attrs) |
---|
90 | { |
---|
91 | switch($elem) |
---|
92 | { |
---|
93 | case 'CHANNEL': |
---|
94 | $GLOBALS['_rss']->depth++; |
---|
95 | $GLOBALS['_rss']->state[$GLOBALS['_rss']->depth] = 'channel'; |
---|
96 | $GLOBALS['_rss']->tmptitle[$GLOBALS['_rss']->depth] = ''; |
---|
97 | $GLOBALS['_rss']->tmplink[$GLOBALS['_rss']->depth] = ''; |
---|
98 | $GLOBALS['_rss']->tmpdesc[$GLOBALS['_rss']->depth] = ''; |
---|
99 | break; |
---|
100 | case 'IMAGE': |
---|
101 | $GLOBALS['_rss']->depth++; |
---|
102 | $GLOBALS['_rss']->state[$GLOBALS['_rss']->depth] = 'image'; |
---|
103 | $GLOBALS['_rss']->tmptitle[$GLOBALS['_rss']->depth] = ''; |
---|
104 | $GLOBALS['_rss']->tmplink[$GLOBALS['_rss']->depth] = ''; |
---|
105 | $GLOBALS['_rss']->tmpdesc[$GLOBALS['_rss']->depth] = ''; |
---|
106 | $GLOBALS['_rss']->tmpurl[$GLOBALS['_rss']->depth] = ''; |
---|
107 | break; |
---|
108 | case 'ITEM': |
---|
109 | $GLOBALS['_rss']->depth++; |
---|
110 | $GLOBALS['_rss']->state[$GLOBALS['_rss']->depth] = 'item'; |
---|
111 | $GLOBALS['_rss']->tmptitle[$GLOBALS['_rss']->depth] = ''; |
---|
112 | $GLOBALS['_rss']->tmplink[$GLOBALS['_rss']->depth] = ''; |
---|
113 | $GLOBALS['_rss']->tmpdesc[$GLOBALS['_rss']->depth] = ''; |
---|
114 | break; |
---|
115 | case 'TITLE': |
---|
116 | $GLOBALS['_rss']->depth++; |
---|
117 | $GLOBALS['_rss']->state[$GLOBALS['_rss']->depth] = 'title'; |
---|
118 | break; |
---|
119 | case 'LINK': |
---|
120 | $GLOBALS['_rss']->depth++; |
---|
121 | $GLOBALS['_rss']->state[$GLOBALS['_rss']->depth] = 'link'; |
---|
122 | break; |
---|
123 | case 'DESCRIPTION': |
---|
124 | $GLOBALS['_rss']->depth++; |
---|
125 | $GLOBALS['_rss']->state[$GLOBALS['_rss']->depth] = 'desc'; |
---|
126 | break; |
---|
127 | case 'URL': |
---|
128 | $GLOBALS['_rss']->depth++; |
---|
129 | $GLOBALS['_rss']->state[$GLOBALS['_rss']->depth] = 'url'; |
---|
130 | break; |
---|
131 | } |
---|
132 | } |
---|
133 | |
---|
134 | function _rssparse_end_elem ($parser, $elem) |
---|
135 | { |
---|
136 | switch ($elem) |
---|
137 | { |
---|
138 | case 'CHANNEL': |
---|
139 | $GLOBALS['_rss']->set_channel( |
---|
140 | $GLOBALS['_rss']->tmptitle[$GLOBALS['_rss']->depth], |
---|
141 | $GLOBALS['_rss']->tmplink[$GLOBALS['_rss']->depth], |
---|
142 | $GLOBALS['_rss']->tmpdesc[$GLOBALS['_rss']->depth] |
---|
143 | ); |
---|
144 | $GLOBALS['_rss']->depth--; |
---|
145 | break; |
---|
146 | case 'IMAGE': |
---|
147 | $GLOBALS['_rss']->set_image( |
---|
148 | $GLOBALS['_rss']->tmptitle[$GLOBALS['_rss']->depth], |
---|
149 | $GLOBALS['_rss']->tmplink[$GLOBALS['_rss']->depth], |
---|
150 | $GLOBALS['_rss']->tmpdesc[$GLOBALS['_rss']->depth], |
---|
151 | $GLOBALS['_rss']->tmpurl[$GLOBALS['_rss']->depth] |
---|
152 | ); |
---|
153 | $GLOBALS['_rss']->depth--; |
---|
154 | break; |
---|
155 | case 'ITEM': |
---|
156 | $GLOBALS['_rss']->add_item( |
---|
157 | $GLOBALS['_rss']->tmptitle[$GLOBALS['_rss']->depth], |
---|
158 | $GLOBALS['_rss']->tmplink[$GLOBALS['_rss']->depth], |
---|
159 | $GLOBALS['_rss']->tmpdesc[$GLOBALS['_rss']->depth] |
---|
160 | ); |
---|
161 | $GLOBALS['_rss']->depth--; |
---|
162 | break; |
---|
163 | case 'TITLE': |
---|
164 | $GLOBALS['_rss']->depth--; |
---|
165 | break; |
---|
166 | case 'LINK': |
---|
167 | $GLOBALS['_rss']->depth--; |
---|
168 | break; |
---|
169 | case 'DESCRIPTION': |
---|
170 | $GLOBALS['_rss']->depth--; |
---|
171 | break; |
---|
172 | case 'URL': |
---|
173 | $GLOBALS['_rss']->depth--; |
---|
174 | break; |
---|
175 | } |
---|
176 | } |
---|
177 | |
---|
178 | function _rssparse_elem_data ($parser, $data) |
---|
179 | { |
---|
180 | switch ($GLOBALS['_rss']->state[$GLOBALS['_rss']->depth]) |
---|
181 | { |
---|
182 | case 'title': |
---|
183 | $GLOBALS['_rss']->tmptitle[($GLOBALS['_rss']->depth - 1)] .= $data; |
---|
184 | break; |
---|
185 | case 'link'; |
---|
186 | $GLOBALS['_rss']->tmplink[($GLOBALS['_rss']->depth - 1)] .= $data; |
---|
187 | break; |
---|
188 | case 'desc': |
---|
189 | $GLOBALS['_rss']->tmpdesc[($GLOBALS['_rss']->depth - 1)] .= $data; |
---|
190 | break; |
---|
191 | case 'url': |
---|
192 | $GLOBALS['_rss']->tmpurl[($GLOBALS['_rss']->depth - 1)] .= $data; |
---|
193 | break; |
---|
194 | } |
---|
195 | } |
---|
196 | |
---|
197 | class rssparser |
---|
198 | { |
---|
199 | var $title; |
---|
200 | var $link; |
---|
201 | var $desc; |
---|
202 | var $items = array(); |
---|
203 | var $nitems; |
---|
204 | var $image = array(); |
---|
205 | var $state = array(); |
---|
206 | var $tmptitle = array(); |
---|
207 | var $tmplink = array(); |
---|
208 | var $tmpdesc = array(); |
---|
209 | var $tmpurl = array(); |
---|
210 | var $depth; |
---|
211 | |
---|
212 | function rssparser() |
---|
213 | { |
---|
214 | $this->nitems = 0; |
---|
215 | $this->depth = 0; |
---|
216 | } |
---|
217 | |
---|
218 | function set_channel($in_title, $in_link, $in_desc) |
---|
219 | { |
---|
220 | $this->title = $in_title; |
---|
221 | $this->link = $in_link; |
---|
222 | $this->desc = $in_desc; |
---|
223 | } |
---|
224 | |
---|
225 | function set_image($in_title, $in_link, $in_desc, $in_url) |
---|
226 | { |
---|
227 | $this->image['title'] = $in_title; |
---|
228 | $this->image['link'] = $in_link; |
---|
229 | $this->image['desc'] = $in_desc; |
---|
230 | $this->image['url'] = $in_url; |
---|
231 | } |
---|
232 | |
---|
233 | function add_item($in_title, $in_link, $in_desc) |
---|
234 | { |
---|
235 | $this->items[$this->nitems]['title'] = $in_title; |
---|
236 | $this->items[$this->nitems]['link'] = $in_link; |
---|
237 | $this->items[$this->nitems]['desc'] = $in_desc; |
---|
238 | $this->nitems++; |
---|
239 | } |
---|
240 | |
---|
241 | function parse($fp) |
---|
242 | { |
---|
243 | $xml_parser = xml_parser_create(); |
---|
244 | |
---|
245 | xml_set_element_handler($xml_parser, '_rssparse_start_elem', '_rssparse_end_elem'); |
---|
246 | xml_set_character_data_handler($xml_parser, '_rssparse_elem_data'); |
---|
247 | |
---|
248 | while ($data = fread($fp, 4096)) |
---|
249 | { |
---|
250 | if (!xml_parse($xml_parser, $data, feof($fp))) |
---|
251 | { |
---|
252 | return 1; |
---|
253 | } |
---|
254 | } |
---|
255 | |
---|
256 | xml_parser_free($xml_parser); |
---|
257 | |
---|
258 | return 0; |
---|
259 | } |
---|
260 | } |
---|
261 | |
---|
262 | function rssparse ($fp) |
---|
263 | { |
---|
264 | $GLOBALS['_rss'] = new rssparser(); |
---|
265 | |
---|
266 | if ($GLOBALS['_rss']->parse($fp)) |
---|
267 | { |
---|
268 | return 0; |
---|
269 | } |
---|
270 | |
---|
271 | return $GLOBALS['_rss']; |
---|
272 | } |
---|
273 | ?> |
---|