fix sitemap
Former-commit-id: ddc69719d5b15a16ac0ec2110cd9fdd9fb44cc0e
--- a/documents/scrapers/f5ce2d1651739704634eb8ca4b2b46d3.py
+++ b/documents/scrapers/f5ce2d1651739704634eb8ca4b2b46d3.py
@@ -7,12 +7,12 @@
#http://www.doughellmann.com/PyMOTW/abc/
class ScraperImplementation(genericScrapers.GenericOAICDisclogScraper):
def getTable(self,soup):
- return soup.find(id = "ctl00_PlaceHolderMain_PublishingPageContent__ControlWrapper_RichHtmlField").table
+ return soup.find(id = "block-system-main").table
def getColumnCount(self):
- return 7
+ return 2
def getColumns(self,columns):
- (id, date, title, description,link,deldate, notes) = columns
- return (id, date, title, description, notes)
+ (date, title) = columns
+ return (date, date, title, title, None)
if __name__ == '__main__':
print 'Subclass:', issubclass(ScraperImplementation, genericScrapers.GenericOAICDisclogScraper)
--- a/documents/sitemap.xml.php
+++ b/documents/sitemap.xml.php
@@ -1,30 +1,48 @@
<?php
include ('../include/common.inc.php');
-$last_updated = date('Y-m-d', @filemtime('cbrfeed.zip'));
header("Content-Type: text/xml");
echo "<?xml version='1.0' encoding='UTF-8'?>";
echo '<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">' . "\n";
echo " <url><loc>" . local_url() . "index.php</loc><priority>1.0</priority></url>\n";
foreach (scandir("./") as $file) {
- if (strpos($file, ".php") !== false && $file != "index.php" && $file != "sitemap.xml.php")
+ if (strpos($file, ".php") !== false && $file != "index.php" && $file != "sitemap.xml.php") {
echo " <url><loc>" . local_url() . "$file</loc><priority>0.6</priority></url>\n";
+ }
}
$agenciesdb = $server->get_db('disclosr-agencies');
+$foidocsdb = $server->get_db('disclosr-foidocuments');
try {
$rows = $agenciesdb->get_view("app", "byCanonicalName")->rows;
foreach ($rows as $row) {
echo '<url><loc>' . local_url() . 'agency.php?id=' . $row->value->_id . "</loc><priority>0.3</priority></url>\n";
}
+ unset($rows);
+ $rows = null;
} catch (SetteeRestClientException $e) {
setteErrorHandler($e);
}
-$foidocsdb = $server->get_db('disclosr-foidocuments');
+
+foreach (range(0, 8) as $number) {
try {
- $rows = $foidocsdb->get_view("app", "all")->rows;
+ $rows = $foidocsdb->get_view("app", "all", Array($number,$number+1))->rows;
foreach ($rows as $row) {
echo '<url><loc>' . local_url() . 'view.php?id=' . $row->value->_id . "</loc><priority>0.3</priority></url>\n";
}
+ unset($rows);
+ $rows = null;
+} catch (SetteeRestClientException $e) {
+ setteErrorHandler($e);
+}
+}
+
+try {
+ $rows = $foidocsdb->get_view("app", "all", Array('9','fffffffff'))->rows;
+ foreach ($rows as $row) {
+ echo '<url><loc>' . local_url() . 'view.php?id=' . $row->value->_id . "</loc><priority>0.3</priority></url>\n";
+ }
+ unset($rows);
+ $rows = null;
} catch (SetteeRestClientException $e) {
setteErrorHandler($e);
}
--- a/include/common.inc.php
+++ b/include/common.inc.php
@@ -10,6 +10,7 @@
|| strstr($_SERVER['PHP_SELF'], "documents/")
|| $_SERVER['SERVER_NAME'] == "disclosurelo.gs"
|| $_SERVER['SERVER_NAME'] == "www.disclosurelo.gs"
+ || $_SERVER['SERVER_NAME'] == "direct.disclosurelo.gs"
)
$basePath = "../";