| 1 | <?php if (!defined('PmWiki')) exit(); |
| 2 | /* |
| 3 | $Id: sitemap.php,v 1.7 2005/12/29 10:26:50 pts00065 Exp $ |
| 4 | This file is NOT part of PmWiki; still you can redistribute it and/or modify |
| 5 | it under the terms of the GNU General Public License as published |
| 6 | by the Free Software Foundation; either version 2 of the License, or |
| 7 | (at your option) any later version. |
| 8 | |
| 9 | |
| 10 | This script generated a google sitemap, either automaticly or as action=sitemap or both. |
| 11 | |
| 12 | The automtic version creates a .gz file and is more efficient compared to the action version. |
| 13 | |
| 14 | google requires that the generated file is located in the root of your pmwiki installation, typically the same place as your pmwiki.php or equivalent file. Thus: |
| 15 | |
| 16 | http://www.brambring.nl/pmwiki.php -> sitemap must be in: http://www.brambring.nl/sitemap.xml.gz |
| 17 | |
| 18 | place the file in the cookbook or local directory |
| 19 | and include the file from your farmconfig.php or config.php |
| 20 | include_once("$FarmD/cookbook/sitemap.php"); |
| 21 | |
| 22 | |
| 23 | Then add the URL to your sitemap in the google webform or ping google |
| 24 | |
| 25 | automatic google ping is not implemented yet. |
| 26 | |
| 27 | regards |
| 28 | |
| 29 | bram |
| 30 | |
| 31 | http://www.brambring.nl |
| 32 | |
| 33 | $Log: sitemap.php,v $ |
| 34 | Revision 1.7 2005/12/29 10:26:50 pts00065 |
| 35 | * support EnablePageListProtect |
| 36 | * Added Site to exclude pattern |
| 37 | |
| 38 | |
| 39 | |
| 40 | */ |
| 41 | SDV($LastModFile, "$WorkDir/.lastmod"); # same as in caching |
| 42 | SDV($SitemapFile, "sitemap.xml.gz"); #will need write access must be in root dir. Ensure dir is writable or create (symbolic) link |
| 43 | |
| 44 | SDV($HandleActions['sitemap'], 'HandleSitemap'); # it is not usefull to have both an action |
| 45 | // and automatic creation ( SitemapDelay >= 0 |
| 46 | SDV($SitemapDelay, 3600); # Seconds to wait after last edit set to -1 to disable automatic generation |
| 47 | SDV($SitemapSquelch, 12*3600); # Squelch between generations of sitemap |
| 48 | |
| 49 | $RobotActions['sitemap'] = 1; |
| 50 | |
| 51 | SDVA($SitemapSearchPatterns, array()); |
| 52 | $SitemapSearchPatterns[] = '!\.(All)?Recent(Changes|Uploads|Pages)$!'; |
| 53 | $SitemapSearchPatterns[] = '!\.Group(Print)?(Header|Footer|Attributes)$!'; |
| 54 | $SitemapSearchPatterns[] = '!^PmWiki\.!'; |
| 55 | $SitemapSearchPatterns[] = '!^Site\.!'; |
| 56 | $SitemapSearchPatterns[] = '!\.SideBar!'; |
| 57 | |
| 58 | SDV($SitemapMaxItems, 50000); # maximum items to display defined by google |
| 59 | SDV($SitemapMaxSize, 10); # maximum size is 10 Mbytes TODO |
| 60 | SDV($SitemapPing, "http://www.google.com/"); # Use ping with long SitemapDelay (like 24*60*60 ) TODO |
| 61 | |
| 62 | // SDV($SitemapTimeFmt,'%Y-%m-%dT%H:%M:%sZ'); # seems to break in current version of google |
| 63 | SDV($SitemapTimeFmt, '%Y-%m-%d'); |
| 64 | |
| 65 | SDV($SiteMapItems, array()); |
| 66 | SDV($SitemapChannelFmt, '<?xml version="1.0" encoding="UTF-8"?> |
| 67 | <urlset xmlns="http://www.google.com/schemas/sitemap/0.84"> |
| 68 | '); |
| 69 | SDV($SitemapItemFmt, ' |
| 70 | <url> |
| 71 | <loc>$PageUrl</loc> |
| 72 | <lastmod>$SitemapItemPubDate</lastmod> |
| 73 | <changefreq>$SitemapChangeFreq</changefreq> |
| 74 | <priority>$SitemapPriority</priority> |
| 75 | </url>'); |
| 76 | SDV($HandleSitemapFmt, array(&$SitemapChannelFmt, &$SitemapItems, '</urlset>')); |
| 77 | |
| 78 | if ( $action == 'browse' ) { |
| 79 | if ($SitemapDelay >= 0) { |
| 80 | $l = @filemtime($LastModFile); |
| 81 | $s = @filemtime($SitemapFile); |
| 82 | if ((($Now - $l) > $SitemapDelay) && ($l > $s) && (($Now - $s) > $SitemapSquelch)) { |
| 83 | $fp = @fopen($SitemapFile, "w"); |
| 84 | if ($fp) { |
| 85 | ob_start(); |
| 86 | MakeSitemap(); |
| 87 | $x = gzencode (ob_get_clean(), 9); |
| 88 | fwrite($fp, $x); |
| 89 | fclose($fp); |
| 90 | } |
| 91 | } |
| 92 | } |
| 93 | } |
| 94 | |
| 95 | |
| 96 | |
| 97 | function HandleSitemap() |
| 98 | { |
| 99 | header("Content-type: text/xml"); |
| 100 | MakeSitemap(); |
| 101 | exit; |
| 102 | } |
| 103 | |
| 104 | function MakeSitemap() |
| 105 | { |
| 106 | global $SitemapMaxItems, $SitemapChannelFmt, $SitemapTimeFmt, |
| 107 | $SitemapItems, $SitemapItemFmt, $SearchPatterns,$FarmD, |
| 108 | $EnablePageListProtect, |
| 109 | $HandleSitemapFmt, $FmtV, $SitemapSearchPatterns, $Now; |
| 110 | global $EntitiesTable; |
| 111 | if (IsEnabled($EnablePageListProtect, 1)) $readf = 1000; |
| 112 | |
| 113 | $t = array(); |
| 114 | $t = @ListPages($SitemapSearchPatterns); |
| 115 | $daily_weekly = 60 * 60 * 24 * 6; #TODO |
| 116 | foreach ($t as $i => $pn) { |
| 117 | $page= ($readf >= 1000) |
| 118 | ? RetrieveAuthPage($pn, 'read', false, READPAGE_CURRENT) |
| 119 | : ReadPage($pn, READPAGE_CURRENT); |
| 120 | if (!$page) continue; |
| 121 | |
| 122 | // foreach ( $page as $k => $l ) { print "$k == $l <br />\n"; } |
| 123 | if ( (count($SitemapItems) > $SitemapMaxItems)) continue; |
| 124 | $FmtV['$SitemapChangeFreq'] = ($Now - $page['time'] < $daily_weekly)?'daily':'weekly'; #TODO |
| 125 | $FmtV['$SitemapPriority'] = '0.5'; #TODO |
| 126 | $FmtV['$SitemapItemPubDate'] = gmstrftime($SitemapTimeFmt, $page['time']); |
| 127 | $SitemapItems[] = FmtPageName($SitemapItemFmt, $page['name']); |
| 128 | } |
| 129 | |
| 130 | #PrintFmt('', str_replace(array_keys($EntitiesTable), array_values($EntitiesTable), $HandleSitemapFmt)); |
| 131 | PrintFmt('', $HandleSitemapFmt); |
| 132 | } |
| 133 | |
| 134 | |