--- a/admin/importAusbudget.php +++ b/admin/importAusbudget.php @@ -1,51 +1,69 @@ get_db('disclosr-agencies'); +$rows = $db->get_view("app", "byName")->rows; +$nametoid = Array(); +$accounts = Array(); +foreach ($rows as $row) { + $nametoid[trim($row->key)] = $row->value; +} +$nameField = "Agency"; +$elevenField = "2011-2012"; +$twelveField = "2012-2013"; +// $request = Requests::get($url); +//echo $url; +// $Data = str_getcsv($request->body, "\n"); //parse the rows +$Data = str_getcsv(file_get_contents("ausbudget.csv"), "\n"); //parse the rows +$headers = Array(); +foreach ($Data as $num => $line) { + $Row = str_getcsv($line, ","); + if ($num == 0) { + $headers = $Row; + print_r($headers); + } else { + if (isset($Row[array_search($nameField, $headers)])) { + $agencyName = $Row[array_search($nameField, $headers)]; + if (!in_array(trim($agencyName), array_keys($nametoid))) { + echo trim($agencyName) . " missing" . PHP_EOL; + } else { + //echo $Row[array_search($nameField, $headers)] . PHP_EOL; +// print_r($Row); + if (!is_array($accounts[$nametoid[trim($agencyName)]]["budget"]['2011-2012'])) $accounts[$nametoid[trim($agencyName)]]["budget"]['2011-2012'] = Array(); + if (!is_array($accounts[$nametoid[trim($agencyName)]]["budget"]['2012-2013'])) $accounts[$nametoid[trim($agencyName)]]["budget"]['2012-2013'] = Array(); -function getTextFromNode($Node, $Text = "") { - if (!isset($Node->tagName) || $Node->tagName == null) { - return $Text . $Node->textContent; + if (!isset($accounts[$nametoid[trim($agencyName)]]["budget"]['2011-2012']['value'])) $accounts[$nametoid[trim($agencyName)]]["budget"]['2011-2012']['value'] = 0; + if (!isset($accounts[$nametoid[trim($agencyName)]]["budget"]['2012-2013']['value'])) $accounts[$nametoid[trim($agencyName)]]["budget"]['2012-2013']['value'] = 0; + $accounts[$nametoid[trim($agencyName)]]["budget"]['2011-2012']['value'] += $Row[array_search($elevenField, $headers)]; + $accounts[$nametoid[trim($agencyName)]]["budget"]['2012-2013']['value'] += $Row[array_search($twelveField, $headers)]; + + $accounts[$nametoid[trim($agencyName)]]["budget"]['2011-2012']['source'] = 'http://budgetaus.net/'; + $accounts[$nametoid[trim($agencyName)]]["budget"]['2012-2013']['source'] = 'http://budgetaus.net/'; + } + } else { + echo "error finding any agency" . $line . PHP_EOL; + } } - $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/"; -$dhandle = opendir("./ausbudget/"); -// define an array to hold the files -$files = array(); -if ($dhandle) { - // loop through all of the files - while (false !== ($fname = readdir($dhandle))) { - if (($fname != '.') && ($fname != '..')) { - echo "$fname
"; - $html = phpQuery::newDocumentHTML(file_get_contents($dir . $fname)); - phpQuery::selectDocument($html); - foreach (pq('table')->elements as $table) { - echo "loltable"; - //echo $table->ownerDocument->saveXML($table); - foreach (pq('tr')->elements as $row) { - echo "lolrow"; - $rowText = getTextFromNode($row); - if (strpos($rowText,"Twitter") === false) { - echo "".pq("td:first",$row)->text()."
"; - echo pq("td",$row)->text() . "
"; - } - } - } + + +//print_r($accounts); +foreach ($accounts as $id => $allvalues) { + echo $id . "
" . PHP_EOL; + $doc = object_to_array($db->get($id)); + // print_r($doc); + + foreach ($allvalues as $valueType => $values) { + if (!isset($doc['statistics'][$valueType]) || !is_array($doc['statistics'][$valueType])) { + + $doc['statistics'][$valueType] = Array(); } - break; + $doc['statistics'][$valueType] = array_unique(array_merge($doc['statistics'][$valueType], $values)); } + + $db->save($doc); } ?>