Link amendments made in new numbering scheme
[contractdashboard.git] / admin / partialdata / scraper.php
blob:a/admin/partialdata/scraper.php -> blob:b/admin/partialdata/scraper.php
--- a/admin/partialdata/scraper.php
+++ b/admin/partialdata/scraper.php
@@ -1,1 +1,49 @@
+<?php
+date_default_timezone_set('Australia/Melbourne');
 
+function format_bytes($size) {
+    $units = array(' B', ' KB', ' MB', ' GB', ' TB');
+    for ($i = 0; $size >= 1024 && $i < 4; $i++) $size /= 1024;
+    return round($size, 2).$units[$i];
+}
+
+$days = 4;
+if (isset($_REQUEST['days'])) $days = $_REQUEST['days'];
+$startDate = strtotime("18-Nov-2007");
+if (isset($_REQUEST['startDate'])) $startDate = $_REQUEST['startDate'];
+function getFile($startDate, $days, $minVal, $maxVal) {
+	$endDate = strtotime(date("Y-m-d", $startDate)." +".$days." days");
+$file = date("dMY",$startDate).'to'.date("dMY",$endDate).'val'.$minVal.'to'.$maxVal.'.xls';
+echo "Fetching $file ($days days) ($minVal < value < $maxVal )... ";
+$url = "https://www.tenders.gov.au/?event=public.advancedsearch.CNSONRedirect&type=cnEvent&atmType=archived%2Cclosed%2Cpublished%2Cproposed&agencyUUID=&agencyStatus=-1&portfolioUUID=&keyword=&KeywordTypeSearch=AllWord&CNID=&dateType=Publish+Date&dateStart=".date("d-M-Y",$startDate)."&dateEnd=".date("d-M-Y",$endDate)."&supplierName=&supplierABN=&valueFrom=".$minVal."&valueTo=".$maxVal."&ATMID=&AgencyRefId=&consultancy=&download=Download+results";
+echo "<!-- $url -->";
+$current = file_get_contents($url);
+if (strpos($current,"Your search returned more than 1000 results.") === false) {
+	file_put_contents($file, $current);
+	echo "$file saved<br>";
+	echo format_bytes(filesize($file))."<br>";
+	echo '<a href="?startDate='.$endDate.'&days='.$days.'">Load next '.($days).' days </a><br>';
+		echo '<a href="?startDate='.$endDate.'&days='.($days*2).'">Load next '.($days*2).' days </a><br>';
+	echo '<a href="?startDate='.$endDate.'&days='.$days.'&split=yes">Load next '.($days).' days with split</a><br>';
+	flush();
+	return true;
+} else  {
+	echo "<font color=red>Too many records!</font><br>";
+	echo '<a href="?startDate='.$startDate.'&days='.floor($days/2).'">Load '.($days/2).' days instead?</a><br>';
+		echo '<a href="?startDate='.$startDate.'&days='.$days.'&split=yes">Split instead?</a><br>';
+	flush();
+	return false;
+}
+}
+if (isset($_REQUEST['split'])) {
+	getFile($startDate, $days, 0, 16000);
+	getFile($startDate, $days, 16000, 20000);
+getFile($startDate, $days, 20000, 30000);
+getFile($startDate, $days, 30000, 40000);
+ getFile($startDate, $days, 40000, 80000);
+ getFile($startDate, $days, 80000, 300000);
+ getFile($startDate, $days, 300000, 999999999);
+} else {
+	getFile($startDate, $days, "" , "");
+}
+?>