function array2str($D) { #$D = compact('data','format', 'select','returnFields', 'letter','cols', 'db') extract($D); $select = $s3ql['select'] != '' ? $s3ql['select'] : $select; if (!ereg('json|php|xml|rdf|n3|sif', $format)) { if ($format == 'html.pretty') { $format = 'html'; echo '<script type="text/javascript"> <!-- function paintRows(){ lines = document.getElementsByTagName(\'tr\'); for (i=0; i<lines.length; i=i+2) {lines[i].style.backgroundColor = \'#BBFFFF\'} for (i=1; i<lines.length; i=i+2) {lines[i].style.backgroundColor = \'lightyellow\'}; } //--> </script>'; echo '<body onload = paintRows()>'; } elseif (ereg('html.(.*)', $format, $css)) { $format = 'html'; $style = @stream_get_contents(@fopen($css[1], 'r')); echo '<style type="text/css">'; echo $style; echo '</style>'; } $format = get_parser_characters($format); #Fetch the cols of what is to be returned if ($select != '') { $P['out'] = urldecode($select); $P['SQLfun'] = ereg_replace("\\(.*\\)", "", $P['out']); if ($P['out'] == $P['SQLfun']) { $P['SQLfun'] = ''; } } #After the query, resume the cols that are supposed to be displayed. Remove the sensitivy cols that should not be displayed if ($P['out'] != '' && $P['out'] != '*') { if ($P['SQLfun'] == 'distinct') { $P['out'] = str_replace(array($P['SQLfun'], "(", ")"), "", $P['out']); } $cols = array_map('trimmit', explode(',', $P['out'])); } #echo '<pre>';print_r($cols); $c = 0; foreach ($cols as $i => $name) { $pCol = $name; #if($c==count($cols)-1) if ($c < count($cols) - 1) { $header .= trim($pCol) . $format['middle']; } else { $header .= trim($pCol); } $c++; } $x = array('data' => $data, 'format' => $format, 'header' => $header, 'columns' => $cols); return display($x); #exit; } else { #filter data by selected if ($s3ql['select'] != '*') { #$t=$GLOBALS['s3codes'][$letter]; #$t=$GLOBALS['plurals'][$t]; #$toreplace = array_keys($GLOBALS['s3map'][$t]); #$replacements = array_values($GLOBALS['s3map'][$t]); #$s3ql['select'] = str_replace($toreplace, $replacements, $s3ql['select']); $s3ql_out = ereg_replace(' ', '', $s3ql['select']); #take out all the spaces $selectFields = explode(',', $s3ql_out); } #clean up the non display field first #echo '<pre>';print_r($data); #echo '<pre>';print_r($selectFields); foreach ($data as $key => $value) { if (!empty($selectFields[0])) { foreach ($selectFields as $colname) { $data2display[$key][$colname] = $value[$colname]; } } else { //echo '<pre>';print_r($value); foreach ($cols as $colname) { $data2display[$key][$colname] = $value[$colname]; } } } $data = $data2display; if ($_REQUEST['out'] == 'header' || $format == 'json' && $_SERVER['HTTPS']) { header("Pragma: public"); header("Expires: 0"); // set expiration time header("Cache-Control: must-revalidate, post-check=0, pre-check=0"); header("Content-Type: application/force-download"); header("Content-Type: application/octet-stream"); header("Content-Type: application/download"); #header("Content-Type: ".$ext.""); // use the Content-Disposition header to supply a recommended filename and // force the browser to display the save dialog. header("Content-Disposition: attachment; filename=s3db." . $format . ""); header("Content-Transfer-Encoding: binary"); } if ($format == 'json') { return 's3db_json(' . json_encode($data) . ')'; exit; } elseif ($format == 'php') { return serialize($data); #echo '$data = '; #print_r($data); exit; } elseif ($format == 'xml') { #header("Content-type: application/xml"); return xml_encode($data, $letter); exit; } elseif ($format == 'rdf' || $format == 'n3') { include 'rdfheader.inc.php'; return rdf_encode($data, $letter, $format, $db); exit; } elseif ($format == 'sif') { return tab_encode($data, $returnFields); exit; } } }
function sparql($I) { ##Parse the query and build the dataset #global $timer; if (is_file(S3DB_SERVER_ROOT . '/pearlib/Benchmark/Timer.php')) { require_once S3DB_SERVER_ROOT . '/pearlib/Benchmark/Timer.php'; $timer = new Benchmark_Timer(); $timer->start(); } extract($I); ##To use SPARQL with ARC library, we will need it to work with a remote endpoint. That means that we do not want to configure ARC as a datastore, but rather to retrieve the data from s3db deployments, convert it to RDF and then use ARC to run the query on it /* ARC2 static class inclusion */ ini_set("include_path", S3DB_SERVER_ROOT . "/pearlib/arc" . PATH_SEPARATOR . ini_get("include_path")); include_once "ARC2.php"; $s3ql['url'] = $in['url'] != '' ? $in['url'] : $default_uri; $s3ql['key'] = $in['key'] != '' ? $in['key'] : get_user_key($user_id, $db); $q = $in['query']; list($query, $triples, $prefixes) = parse_sparql_query($q, $s3ql); $bq .= "PREFIX " . implode("\n PREFIX ", $query['prefix']) . "\n "; $bq .= "SELECT " . $query['select'][0] . "\n "; $bq .= "FROM" . implode(" FROM ", $query['from']) . "\n "; $bq .= "WHERE " . $query['where'][0] . "\n "; preg_match_all('(\\?[A-Za-z0-9]+) ', $bq, $vars); if ($vars[0]) { $vars = array_unique($vars[0]); $sparql_vars = implode(" ", $vars); } if ($query['select'][0] != "" && $query['select'][0] != "*") { $outputCols = explode(" ", trim($query['select'][0])); $outputCols = array_filter($outputCols); $outputCols = array_intersect($vars, $outputCols); } $sparql = ereg_replace("FROM(.*)WHERE", "WHERE", $bq); #lets preprocess the order by which the must be queries must be performed to optimize speedness list($iterations, $scrambled) = iterationOrder($triples, $prefixes, true); ##$rdf_results will contain the totality of triples retrieved from s3db; ##Start a rdf-api model $iterations = array_values($iterations); $rdf = S3DB_URI_BASE . '/s3dbcore/model.n3'; #base s3db rdf model $filename = md5($rdf); $file_place = $GLOBALS['uploads'] . '/'; #$queryModel = rdf2php($rdf); #$data = $queryModel->sparqlQuery($sparql); #echo '<pre>';print_r($data);exit; if ($timer) { $timer->setMarker('Core model read into results'); } $rdf_results = array(); $performedQueries = array(); $r = 0; foreach ($iterations as $it => $triples2query) { $S3QL = array(); $S3QLfinal = array(); foreach ($triples2query as $i => $tripleInd) { $tripleString = $tripleInd; list($subject, $predicate, $object) = explode(' ', trim($tripleString)); $subject = ereg_replace('^<|>$', '', $subject); $predicate = ereg_replace('^<|>$', '', $predicate); $object = ereg_replace('^<|>$', '', $object); $triple = compact('subject', 'predicate', 'object'); #sparql triple is used to calculate the values of the variables in the triple #$sparql_triple = $sparql_prefixes_default.' SELECT * WHERE { '.ltrim($tripleString).' . }'; #now lets interpret the triple to explore the space of possible queries on S3QL $pack = compact('triple', 's3ql', 'user_id', 'db', 'prefixes', 'varType', 'discoveredData', 'it', 'varTypeWhere', 'collected_data', 'performedQueries'); $sp = sparql_navigator($pack); extract($sp); # if($timer) $timer->setMarker('Built query '.$i); ##Remove queries that were already performed if ($S3QL[0]) { foreach ($S3QL as $s => $q) { $S3QLfinal[] = $q; $queried_elements[] = $element[$s]; } $localQueries[$tripleString] = $localQueries[0]; $remoteQueries[$tripleString] = $remoteQueries[0]; $localQueries = array_filter($localQueries); $remoteQueries = array_filter($remoteQueries); } } $S3QL = $S3QLfinal; ##Remove repeated queries $S3QL = array_unique($S3QL); #if only the s3ql is requested, we can return it now if ($in['output'] == 'S3QL') { foreach ($localQueries as $sparqlVersion => $s3qlVersion) { $Q[]['S3QL'] = S3QLQuery($s3qlVersion); } foreach ($remoteQueries as $rq) { $Q[]['S3QL'] = $rq; } $root = 's3ql'; #root is just the word that xml should parse as the root for each entry $data = $Q; $cols = array('S3QL'); $format = $in['format'] == '' ? 'html' : $in['format']; $z = compact('data', 'cols', 'format', 'root'); $out = outputFormat($z); return array(true, $out); } #If paralel library is activated, use it for the data. Otherwise use the custom version #$query_answers_file = 'sparql_query_ans'.rand(100,200); $a=fopen($query_answers_file, 'a'); if (!empty($S3QL)) { if (extension_loaded('curl') && $goparallel) { // Create cURL handlers if ($timer) { $timer->setMarker('Starting queries from group ' . $it); } foreach ($S3QL as $k => $url) { $qURL = $url; $ch[$k] = curl_init(); // Set options curl_setopt($ch[$k], CURLOPT_URL, $qURL . '&format=php'); curl_setopt($ch[$k], CURLOPT_RETURNTRANSFER, 1); } $mh = curl_multi_init(); foreach ($S3QL as $k => $url) { curl_multi_add_handle($mh, $ch[$k]); } $running = null; do { curl_multi_exec($mh, $running); if ($timer) { $timer->setMarker('Query ' . $k . ' of group ' . $it . ' executed'); } } while ($running > 0); foreach ($S3QL as $k => $url) { $answer[$k] = curl_multi_getcontent($ch[$k]); if (!empty($answer[$k])) { #@fwrite($a, $answer[$k]); ##This is what takes the longest after the query, can it be replaced? $ans = unserialize($answer[$k]); $letter = $queried_elements[$r][0]; if (empty($ans)) { ##is this query part is not optional, then the result will be null ##TO BE DEVELOPED SOON } else { $rdf_results[$letter][] = $ans; } $r++; ##Add the triples to already existing triples #Line up the answer with the model if ($timer) { $timer->setMarker('Query ' . $it . '=>' . $k . ' converted to php '); } } } curl_multi_close($mh); ####Time count #$time_end = microtime(true); #$time = $time_end - $time_start; #echo "Query took ".$time." seconds\n";exit; ### } else { #Now solve the remaining triples with the constants found in this one if (is_array($localQueries) && !empty($localQueries)) { foreach ($localQueries as $sparql_triple => $s3ql) { $s3ql = array_filter(array_diff_key($s3ql, array('url' => ''))); $answer = localQ($s3ql); if (!empty($answer)) { $rdfanswer = rdf2php($answer); #Line up the answer with the model $queryModel->addModel($rdfanswer); #Now perform the query on the small model to find a constant for the remaining queries #list($data,$discovered, $discoveredData,$queryModel) = executeQuery($queryModel,$sparql_triple,$discovered,$format); } } } if (is_array($remoteQueries) && !empty($remoteQueries)) { foreach ($remoteQueries as $remoteQuery) { $answer = remoteQ($remoteQuery); if (!empty($answer)) { $rdfanswer = rdf2php($answer); #Line up the answer with the model $queryModel->addModel($rdfanswer); #Now perform the query on the small model to find a constant for the remaining queries #list($data,$discovered, $discoveredData,$queryModel) = executeQuery($queryModel,$sparql_triple,$discovered,$format); } } } } } } ##Get the data from the file ##Now, add the dictionary data if ($complete) { include_once S3DB_SERVER_ROOT . '/s3dbcore/dictionary.php'; $s3qlN = compact('user_id', 'db'); $s3qlN['from'] = 'link'; $s3qlN['format'] = 'php'; $links = query_user_dictionaries($s3qlN, $db, $user_id); $links = unserialize($links); $rdf_results['E'][0] = $links; $s3qlN = compact('user_id', 'db'); $s3qlN['from'] = 'namespaces'; $s3qlN['format'] = 'php'; $ns = query_user_dictionaries($s3qlN, $db, $user_id); $ns = unserialize($ns); if ($timer) { $timer->setMarker('Dictionary links retrieved'); } } ##Convert the result into an RDF file $data_triples = array(); if (is_array($rdf_results)) { foreach ($rdf_results as $letter => $results2rdfize) { $dont_skip_core_name = false; $dont_skip_serialized = true; if (ereg('S', $letter)) { $dont_skip_serialized = false; } if (ereg('C|R|P', $letter)) { $dont_skip_core_name = true; } foreach ($results2rdfize as $k => $data) { $tmp_triples = rdf_encode($data, $letter, 'array', $s3ql['db'], $ns, $collected_data, $dont_skip_serialized, $dont_skip_core_name); if (is_array($tmp_triples)) { $data_triples = array_merge($data_triples, $tmp_triples); } } } } if (!empty($data_triples)) { $tmp['ns'] = $prefixes; /* #this one for turtle $parser = ARC2::getComponent('TurtleParser', $a); $index = ARC2::getSimpleIndex($triples, false) ; # false -> non-flat version $rdf_doc = $parser->toTurtle($index,$prefixes); */ $parser = ARC2::getComponent('RDFXMLParser', $tmp); $index = ARC2::getSimpleIndex($data_triples, false); /* false -> non-flat version */ $rdf_doc = $parser->toRDFXML($index, $prefixes); $filename = S3DB_SERVER_ROOT . '/tmp/' . random_string(15) . '.rdf'; $rr = fopen($filename, 'a+'); fwrite($rr, $rdf_doc); fclose($rr); if ($timer) { $timer->setMarker(count($data_triples) . ' triples written to file ' . $filename); } ##The better strategy would be to let the client cpu resolve the query; return the graphs with the rdf so that a sparql on the client can handle it if ($return_file_name) { if (filesize($filename) > 0) { return array(true, $filename); } else { return array(false); } exit; } if ($redirect) { ##And now use an external service ( I gave up with ARC) to parse the query $url2search = str_replace(S3DB_SERVER_ROOT, S3DB_URI_BASE, $filename); ##Giving up on ARC, surrender to sparql.com $remote_endpoint = "http://sparql.org/sparql?query="; $bq = ereg_replace("FROM <.*>", "FROM <" . $url2search . ">", $bq); $bq = urlencode($bq); $remote_endpoint .= $bq . '&default-graph-uri=&stylesheet=/xml-to-html.xsl'; return array(true, $remote_endpoint); } #echo $filename;exit; #And finally perform the query on the model. $queryModel = rdf2php($filename); $format = $in['format'] != '' ? $in['format'] : 'html'; unlink($filename); if ($timer) { $timer->setMarker('Data converted to a model the rdf-api can query'); } if (eregi('^(sparql-xml|sparql-html)$', $format)) { switch ($format) { case 'sparql-xml': $result = $queryModel->sparqlQuery($sparql, 'XML'); break; case 'sparql-html': $result = $queryModel->sparqlQuery($sparql, 'HTML'); if ($_REQUEST['su3d']) { $timer->stop(); $profiling = $timer->getProfiling(); echo "Query took " . $profiling[count($profiling) - 1]['total'] . ' sec'; } break; } if ($result) { return array(true, $result); } else { return false; } } elseif ($format == 'html.form') { $form .= ' <html> <head> </head><body> <form method="GET" action="sparql.php" id="sparqlform"> <h5>Target Deployment(s)</h5> <input type="hidden" name="key" value="' . $s3ql['key'] . '"/> <input type="hidden" name="format" value="' . $_REQUEST['format'] . '"/> <input type = "text" id="url" size = "100%" value="' . $GLOBALS['url'] . '" name="url"> <h5>SPARQL <a href="http://www.w3.org/TR/rdf-sparql-query/" target="_blank">(help!!)</a></h5> <br /> <textarea cols="100" id="sparql" rows="10" name = "query">' . stripslashes($sparql) . '</textarea><br /> <input type="submit" value="SPARQL this!" id="submitsparql"></body> </form> '; $form .= '<br />' . count($data) . " rows"; $form .= '<br />Query took ' . (strtotime(date('His')) - $start) . ' sec'; if (count($data) > 0) { return array(true, $form); } else { return array(false); } } else { #and output the result according to requested format $data = $queryModel->sparqlQuery($sparql); if ($timer) { $timer->setMarker('Query on SPARQL data executed by rdf-api.'); } if (is_array($outputCols) && !empty($outputCols)) { ##only this one are to be shown in the final result $vars = $outputCols; } $cleanCols = array(); foreach ($vars as $varname) { $cleanCols[] = ereg_replace('^\\?', '', $varname); } $outputData = array(); if (is_array($data)) { foreach ($data as $s => $sparql_line) { foreach ($sparql_line as $sparql_var => $sparql_var_value) { if ($sparql_var_value->uri != '') { $outputData[$s][ereg_replace('^\\?', '', $sparql_var)] = $sparql_var_value->uri; } elseif ($sparql_var_value->label != '') { $outputData[$s][ereg_replace('^\\?', '', $sparql_var)] = $sparql_var_value->label; } else { $outputData[$s][ereg_replace('^\\?', '', $sparql_var)] = ""; } } } } if ($timer) { $timer->setMarker('Data converted in a format that fun outputformat can read'); } #$timer ->display(); #root is just the word that xml should parse as the root for each entry $root = 'sparql'; if ($timer) { $timer->setMarker('All variables fitted into their places to represent in the final output'); } $data = $outputData; $cols = $cleanCols; if ($_REQUEST['su3d']) { $timer->stop(); $profiling = $timer->getProfiling(); echo "Query took " . $profiling[count($profiling) - 1]['total'] . ' sec<br>'; } $z = compact('data', 'cols', 'format', 'root'); $out = outputFormat($z); echo $out; exit; if (count($data) > 0) { return array(true, $out); } else { return array(false); } } } else { return array(false); } #else { #$out= formatReturn($GLOBALS['error_codes']['no_results'], 'Your query did not return any results.', $format,''); #} }
function class_triples($triples, $root, $specified_id, $specified_id_info, $s3Types, $user_id, $db, $inputs, $s3db = array(), $fork = 0, $nest = 0) { $s3idNames = $GLOBALS['COREids']; ##this is the else, but since we re returninng... foreach ($s3Types[$root] as $k => $a_class) { $uid_p = letter($specified_id) . $specified_id_info[$specified_id]; #each class has a descriptive statement $a_class_id = $s3idNames[$a_class]; $a_class_letter = strtoupper(substr($a_class, 0, 1)); $a_class_type = $GLOBALS['s3codes'][$a_class_letter]; $s3ql = compact('user_id', 'db'); $s3ql['select'] = '*'; $s3ql['from'] = $GLOBALS['plurals'][$a_class]; $s3ql['where'][$specified_id] = $specified_id_info[$specified_id]; if (ereg('(rule|statement)', $a_class_type)) { $s3ql['where']['object'] = "!=UID"; } $subClasses = S3QLaction($s3ql); #find them, output them. #$verbs=array(); $s3db[$uid_p][letter($a_class)] = $subClasses; #triples for teh rdf api $subClassTriples = rdf_encode($subClasses, letter($GLOBALS['plurals'][$a_class]), 'array', $db, $namespaces, $subClasses); if (!empty($subClassTriples)) { $triples = array_merge($triples, $subClassTriples); } if (is_array($subClasses) && is_array($s3Types[$a_class])) { ##prepare triples for the rdf-api foreach ($subClasses as $subSub => $subSubInfo) { if ($inputs['all'] == 1 || $a_class == 'collection' && $subSubInfo['name'] == 's3dbVerb') { $triples = class_triples($triples, $a_class, $a_class_id, $subSubInfo, $s3Types, $user_id, $db, $s3db, $fork, $nest = 1); } } } $nest = 0; $fork++; } return $triples; }