| <?php | <?php |
| include_once("../include/common.inc.php"); | include_once("../include/common.inc.php"); |
| require($basePath . 'lib/phpquery/phpQuery/phpQuery.php'); | require($basePath . 'lib/phpquery/phpQuery/phpQuery.php'); |
| setlocale(LC_CTYPE, 'C'); | setlocale(LC_CTYPE, 'C'); |
| function getTextFromNode($Node, $Text = "") { | |
| if (!isset($Node->tagName) || $Node->tagName == null) { | |
| return $Text . $Node->textContent; | |
| } | |
| $Node = $Node->firstChild; | |
| if (isset($Node) && $Node != null) { | |
| $Text = getTextFromNode($Node, $Text); | |
| } | |
| while (isset($Node->nextSibling) && $Node->nextSibling != null) { | |
| $Text = getTextFromNode($Node->nextSibling, $Text); | |
| $Node = $Node->nextSibling; | |
| } | |
| return $Text; | |
| } | |
| $dir = "./ausbudget/"; | $dir = "./ausbudget/"; |
| $dhandle = opendir("./ausbudget/"); | $dhandle = opendir("./ausbudget/"); |
| // define an array to hold the files | $headers = Array("Table ID", "Portfolio", "Agency", "Program", "Scheme", "2011-2012", "2012-2013", "Difference", "Source"); |
| $files = array(); | |
| if ($dhandle) { | $fp = fopen('php://output', 'w'); |
| // loop through all of the files | if ($fp) { |
| while (false !== ($fname = readdir($dhandle))) { | header('Content-Type: text/csv; charset=utf-8'); |
| if (($fname != '.') && ($fname != '..')) { | header('Content-Disposition: attachment; filename="export.' . date("c") . '.csv"'); |
| echo "$fname <br>"; | header('Pragma: no-cache'); |
| $html = phpQuery::newDocumentHTML(file_get_contents($dir . $fname)); | header('Expires: 0'); |
| phpQuery::selectDocument($html); | fputcsv($fp, $headers); |
| foreach (pq('table')->elements as $table) { | if ($dhandle) { |
| echo "loltable"; | // loop through all of the files |
| //echo $table->ownerDocument->saveXML($table); | while (false !== ($fname = readdir($dhandle))) { |
| foreach (pq('tr')->elements as $row) { | if (($fname != '.') && ($fname != '..')) { |
| echo "lolrow"; | //echo "$fname <br>"; |
| $rowText = getTextFromNode($row); | $html = phpQuery::newDocumentHTML(file_get_contents($dir . $fname)); |
| if (strpos($rowText,"Twitter") === false) { | phpQuery::selectDocument($html); |
| echo "<b>".pq("td:first",$row)->text()."</b><br>"; | foreach (pq('table')->elements as $table) { |
| echo pq("td",$row)->text() . "<br>"; | $data = Array(); |
| ////echo "loltable"; | |
| //echo $table->ownerDocument->saveXML($table); | |
| foreach (pq('tr',$table)->elements as $row) { | |
| //echo "lolrow"; | |
| $rowText = pq($row)->text(); | |
| if (strpos($rowText, "Twitter") === false) { | |
| $key = trim(pq("td:first", $row)->text()); | |
| //echo "<b>$key</b><br>"; | |
| $value = trim(str_replace(pq("td:first", $row)->text(), "", pq("td", $row)->text())); | |
| if ($key == "2011-2012" || $key == "2012-2013") { | |
| $eValue = explode("-", $value); | |
| $value = trim(str_replace(Array("$", ",", "\n"), "", $eValue[0])); | |
| } | |
| //echo "$value <br>"; | |
| $data[$key] = trim(str_replace(Array("\n"), "", $value)); | |
| if ($key == "Source") { | |
| foreach ($headers as $fieldName) { | |
| if (isset($data[$fieldName])) { | |
| $csvrow[] = $data[$fieldName]; | |
| } else { | |
| $csvrow[] = ""; | |
| } | |
| } | |
| fputcsv($fp, array_values($csvrow)); | |
| $data = Array(); | |
| $csvrow = Array(); | |
| } | |
| } | |
| } | } |
| } | } |
| } | } |
| } | } |
| break; | |
| } | } |
| die; | |
| } | } |
| ?> | ?> |