I'm using PHP to translate a .gexf output to JSON to look into representation of Gephi output in Protovis/D3 but I'm having a bear of a time accessing the viz: attributes of a node. For some reason, when I check the children of node, I see "attvalues" but I don't see viz:size, viz:position or viz:color. I'm just using simplexml_load_file so if someone has a better process or a built parser, let me know.
Of course, the real thing to do is build a Protovis/D3 exporter (unfortunately, it doesn't parse id values for source/target but instead treats the id value as the order of the node in the JSON) which I suppose I should do.
[SOLVED] Parsing GEXF in PHP
Re: Parsing GEXF in PHP
I had a similar need and I used DOMXPath that works well. You have to register the namespaces with the xPath object:
After that, queries give you all the attributes you want, but you have to explicit the namespaces:
Here is all my code (stores the gexf in a database). I don't claim it's clean or well done, but it works and could help you...
NB: dbh is the database handler.
Code: Select all
$xp->registerNamespace('gexf', 'http://www.gexf.net/1.1draft');
$xp->registerNamespace('viz', 'http:///www.gexf.net/1.1draft/viz');
Code: Select all
$nodesattributes = $xp->query("//gexf:graph[1]/gexf:attributes[@class='node']/gexf:attribute");NB: dbh is the database handler.
Code: Select all
<?php
# GEXF
$gexfDom = new DomDocument();
$gexfDom->load($file->get('path'));
$xp = new DOMXPath($gexfDom);
if($gexfDom->getElementsByTagName('gexf')->item(0)->getAttribute('xmlns') == "http://www.gephi.org/gexf"){
# 'Old style' GEXF
$xp->registerNamespace('gexf', 'http://www.gephi.org/gexf');
$xp->registerNamespace('viz', 'http://www.gephi.org/gexf/viz');
$this->dbh->beginTransaction();
# Nodes table
$nodesQuery = "CREATE TABLE nodes (id INTEGER PRIMARY KEY AUTOINCREMENT, importId TEXT, label TEXT, x REAL, y REAL, size REAL";
# Nodes Attributes
$nodesattributes = $xp->query("//gexf:graph[1]/gexf:attributes[@class='node']/gexf:attribute");
$existingNames = array("id", "importId", "label", "x", "y", "size");
foreach($nodesattributes as $attribute){
$name = $attribute->getAttribute('id');
$name = preg_replace('/[^a-z0-9_]*/i','',$name);
while(in_array($name, $existingNames)){
$name = "old_$name";
}
$existingNames[] = $name;
$name = $this->dbh->quote($name);
$title = $this->dbh->quote($attribute->getAttribute('title'));
$default = $this->dbh->quote($attribute->getAttribute('default'));
$truetype = strtolower($attribute->getAttribute('type'));
if($truetype == "integer" || $truetype == "long"){
$type = "INTEGER";
$nodesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "double" || $truetype == "float"){
$type = "REAL";
$nodesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "boolean" || $truetype == "string" || $truetype == "anyuri"){
$type = "TEXT";
$nodesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "liststring"){
$type = "TAGS";
} else {
die("Error (graph): unsupported attribute type ".strtolower($attribute->getAttribute('type')));
}
$truetype = $this->dbh->quote($truetype);
$type = $this->dbh->quote($type);
$this->dbh->exec(utf8_decode("INSERT INTO nodesattributes ('name', 'title', 'truetype', 'type', 'defaultvalue', 'viztype') VALUES ($name, $title, $truetype, $type, $default, 'unpublished')"));
}
$nodesQuery .= ")";
$this->dbh->exec(utf8_decode($nodesQuery));
$this->dbh->commit();
$this->dbh->beginTransaction();
# Edges table
$edgesQuery = "CREATE TABLE edges (id INTEGER PRIMARY KEY AUTOINCREMENT, importId TEXT, source INTEGER, target INTEGER";
# Edges Attributes
$edgesattributes = $xp->query("//gexf:graph[1]/gexf:attributes[@class='edge']/gexf:attribute");
$existingNames = array("id", "importId", "source", "target");
foreach($edgesattributes as $attribute){
$name = $attribute->getAttribute('id');
$name = preg_replace('/[^a-z0-9_]*/i','',$name);
while(in_array($name, $existingNames)){
$name = "old_$name";
}
$existingNames[] = $name;
$name = $this->dbh->quote($name);
$title = $this->dbh->quote($attribute->getAttribute('title'));
$default = $this->dbh->quote($attribute->getAttribute('default'));
$truetype = strtolower($attribute->getAttribute('type'));
if($truetype == "integer" || $truetype == "long"){
$type = "INTEGER";
$edgesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "double" || $truetype == "float"){
$type = "REAL";
$edgesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "boolean" || $truetype == "string" || $truetype == "anyuri"){
$type = "TEXT";
$edgesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "liststring"){
$type = "TAGS";
} else {
die("Error (graph): unsupported attribute type ".strtolower($attribute->getAttribute('type')));
}
$truetype = $this->dbh->quote($truetype);
$type = $this->dbh->quote($type);
$this->dbh->exec(utf8_decode("INSERT INTO 'edgesattributes' ('name', 'title', 'truetype', 'type', 'defaultvalue', 'viztype') VALUES ($name, $title, $truetype, $type, $default, 'unpublished')"));
}
$edgesQuery .= ")";
$this->dbh->exec(utf8_decode($edgesQuery));
$this->dbh->commit();
$this->loadAttributes();
$tagQueries = array();
# Fill nodes table
$this->dbh->beginTransaction();
$nodes = $xp->query("//gexf:graph[1]/gexf:nodes/gexf:node");
foreach($nodes as $node){
$query_insert = "INSERT INTO nodes ('importid', 'label', 'x', 'y', 'size'";
$importid = $this->dbh->quote($node->getAttribute('id'));
$label = $this->dbh->quote($node->getAttribute('label'));
$x = rand(-100, 100);
$y = rand(-100, 100);
$size = 1;
if($positionElement = $xp->query("//gexf:graph[1]/gexf:nodes/gexf:node[@id='".$node->getAttribute('id')."']/viz:position[1]")){
$x = $this->dbh->quote($positionElement->item(0)->getAttribute('x'));
$y = $this->dbh->quote($positionElement->item(0)->getAttribute('y'));
}
if($sizeElement = $xp->query("//gexf:graph[1]/gexf:nodes/gexf:node[@id='".$node->getAttribute('id')."']/viz:size[1]")){
$size = $this->dbh->quote($sizeElement->item(0)->getAttribute('value'));
}
$query_values = "VALUES ($importid, $label, $x, $y, $size";
if($node->getElementsByTagName('attvalues')->length > 0 && $node->getElementsByTagName('attvalues')->item(0)->getElementsByTagName('attvalue')->length > 0){
foreach($node->getElementsByTagName('attvalues')->item(0)->getElementsByTagName('attvalue') as $value){
$attId = $value->getAttribute('id') + 1;
$attValue = $this->dbh->quote($value->getAttribute('value'));
$att = $this->getNodeAttributeById($attId);
$attName = $this->dbh->quote($att['name']);
if($att['type'] == 'TAGS'){
$tags = explode("|", $value->getAttribute('value'));
foreach($tags as $tag){
$tag = trim($tag);
if($tag != ""){
$tag = $this->dbh->quote($tag);
$tagQueries[] = "INSERT INTO nodestags (attribute, node, value) SELECT $attId, n.id, $tag FROM nodes n WHERE n.importId=$importid";
}
}
} else {
$query_insert .= ", $attName";
$query_values .= ", $attValue";
}
}
}
$this->dbh->exec(utf8_decode("$query_insert ) $query_values )"));
}
$this->dbh->commit();
# Fill edges table
$this->dbh->beginTransaction();
$edges = $xp->query("//gexf:graph[1]/gexf:edges/gexf:edge");
foreach($edges as $edge){
$query_insert = "INSERT INTO 'edges' ('importid', 'source', 'target'";
$importid = $this->dbh->quote($edge->getAttribute('id'));
$source = $this->dbh->quote($edge->getAttribute('source'));
$target = $this->dbh->quote($edge->getAttribute('target'));
$query_select = "SELECT $importid, sourcenodes.id, targetnodes.id";
$query_from = " FROM nodes sourcenodes, nodes targetnodes WHERE sourcenodes.importId=$source AND targetnodes.importId=$target";
if($edge->getElementsByTagName('attvalues')->length > 0 && $edge->getElementsByTagName('attvalues')->item(0)->getElementsByTagName('attvalue')->length > 0){
foreach($edge->getElementsByTagName('attvalues')->item(0)->getElementsByTagName('attvalue') as $value){
$attId = $value->getAttribute('id') + 1;
$attValue = $this->dbh->quote($value->getAttribute('value'));
$att = $this->getNodeAttributeById($attId);
$attName = $this->dbh->quote($att['name']);
if($att['type'] == 'TAGS'){
$tags = explode("|", $value->getAttribute('value'));
foreach($tags as $tag){
$tag = trim($tag);
if($tag != ""){
$tag = $this->dbh->quote($tag);
$tagQueries[] = "INSERT INTO 'edgestags' (attribute, edge, value) SELECT $attId, e.id, $tag FROM edges e WHERE e.importId=$importid";
}
}
} else {
$query_insert .= ", $attName";
$query_values .= ", $attValue";
}
}
}
$this->dbh->exec(utf8_decode("$query_insert ) $query_select $query_from"));
}
$this->dbh->commit();
# Execute the tag queries
$this->dbh->beginTransaction();
foreach($tagQueries as $query){
$this->dbh->exec(utf8_decode($query));
}
$this->dbh->commit();
} else if($gexfDom->getElementsByTagName('gexf')->item(0)->getAttribute('xmlns') == "http://www.gexf.net/1.1draft"){
# 1.1 draft GEXF
$xp->registerNamespace('gexf', 'http://www.gexf.net/1.1draft');
$xp->registerNamespace('viz', 'http:///www.gexf.net/1.1draft/viz');
$this->dbh->beginTransaction();
# Nodes table
$nodesQuery = "CREATE TABLE nodes (id INTEGER PRIMARY KEY AUTOINCREMENT, importid TEXT, label TEXT, x REAL, y REAL, size REAL";
# Nodes Attributes
$nodesattributes = $xp->query("//gexf:graph[1]/gexf:attributes[@class='node']/gexf:attribute");
$existingNames = array("id", "importid", "label", "x", "y", "size");
foreach($nodesattributes as $attribute){
$name = "attr_".$attribute->getAttribute('id');
$name = preg_replace('/[^a-z0-9_]*/i','',$name);
while(in_array($name, $existingNames)){
$name = "old_$name";
}
$existingNames[] = $name;
$name = $this->dbh->quote($name);
$title = $this->dbh->quote($attribute->getAttribute('title'));
$default = $this->dbh->quote($attribute->getAttribute('default'));
$truetype = strtolower($attribute->getAttribute('type'));
if($truetype == "int" || $truetype == "integer" || $truetype == "long"){
$type = "INTEGER";
$nodesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "double" || $truetype == "float"){
$type = "REAL";
$nodesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "boolean" || $truetype == "string" || $truetype == "anyuri"){
$type = "TEXT";
$nodesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "liststring"){
$type = "TAGS";
} else {
die("Error (graph): unsupported attribute type ".strtolower($attribute->getAttribute('type')));
}
$truetype = $this->dbh->quote($truetype);
$type = $this->dbh->quote($type);
$this->dbh->exec(utf8_decode("INSERT INTO nodesattributes ('name', 'title', 'truetype', 'type', 'defaultvalue', 'viztype') VALUES ($name, $title, $truetype, $type, $default, 'unpublished')"));
}
$nodesQuery .= ")";
$this->dbh->exec(utf8_decode($nodesQuery));
$this->dbh->commit();
$this->dbh->beginTransaction();
# Edges table
$edgesQuery = "CREATE TABLE edges (id INTEGER PRIMARY KEY AUTOINCREMENT, importid TEXT, source INTEGER, target INTEGER";
# Edges Attributes
$edgesattributes = $xp->query("//gexf:graph[1]/gexf:attributes[@class='edge']/gexf:attribute");
$existingNames = array("id", "importid", "source", "target", "weight");
foreach($edgesattributes as $attribute){
$name = "attr_".$attribute->getAttribute('id');
$name = preg_replace('/[^a-z0-9_]*/i','',$name);
while(in_array($name, $existingNames)){
$name = "old_$name";
}
$existingNames[] = $name;
$name = $this->dbh->quote($name);
$title = $this->dbh->quote($attribute->getAttribute('title'));
$default = $this->dbh->quote($attribute->getAttribute('default'));
$truetype = strtolower($attribute->getAttribute('type'));
if($truetype == "int" || $truetype == "integer" || $truetype == "long"){
$type = "INTEGER";
$edgesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "double" || $truetype == "float"){
$type = "REAL";
$edgesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "boolean" || $truetype == "string" || $truetype == "anyuri"){
$type = "TEXT";
$edgesQuery .= ", $name $type DEFAULT $default";
} else if($truetype == "liststring"){
$type = "TAGS";
} else {
die("Error (graph): unsupported attribute type ".strtolower($attribute->getAttribute('type')));
}
$truetype = $this->dbh->quote($truetype);
$type = $this->dbh->quote($type);
$this->dbh->exec(utf8_decode("INSERT INTO 'edgesattributes' ('name', 'title', 'truetype', 'type', 'defaultvalue', 'viztype') VALUES ($name, $title, $truetype, $type, $default, 'unpublished')"));
}
$edgesQuery .= ")";
$this->dbh->exec(utf8_decode($edgesQuery));
$this->dbh->commit();
$this->loadAttributes();
$tagQueries = array();
# Fill nodes table
$this->dbh->beginTransaction();
$nodes = $xp->query("//gexf:graph[1]/gexf:nodes/gexf:node");
foreach($nodes as $node){
$query_insert = "INSERT INTO nodes ('importid', 'label', 'x', 'y', 'size'";
$importid = $this->dbh->quote($node->getAttribute('id'));
$label = $this->dbh->quote($node->getAttribute('label'));
$x = rand(-100, 100);
$y = rand(-100, 100);
$size = 1;
if($positionElement = $xp->query("//gexf:graph[1]/gexf:nodes/gexf:node[@id='".$node->getAttribute('id')."']/viz:position[1]")){
$x = $this->dbh->quote($positionElement->item(0)->getAttribute('x'));
$y = $this->dbh->quote($positionElement->item(0)->getAttribute('y'));
}
if($sizeElement = $xp->query("//gexf:graph[1]/gexf:nodes/gexf:node[@id='".$node->getAttribute('id')."']/viz:size[1]")){
$size = $this->dbh->quote($sizeElement->item(0)->getAttribute('value'));
}
$query_values = "VALUES ($importid, $label, $x, $y, $size";
foreach($node->getElementsByTagName('attvalue') as $value){
$attName = "attr_".preg_replace('/[^a-z0-9_]*/i','',$value->getAttribute('for'));
$attValue = $this->dbh->quote($value->getAttribute('value'));
$att = $this->getNodeAttributeByName($attName);
$attName = $this->dbh->quote($att['name']);
if($att['type'] == 'TAGS'){
$tags = explode("|", $value->getAttribute('value'));
foreach($tags as $tag){
$tag = $this->dbh->quote($tag);
$tagQueries[] = "INSERT INTO nodestags (attribute, node, value) SELECT a.id, n.id, $tag FROM nodes n, nodesattributes a WHERE n.importid=$importid AND a.name=$attName";
}
} else {
$query_insert .= ", $attName";
$query_values .= ", $attValue";
}
}
$this->dbh->exec(utf8_decode("$query_insert ) $query_values )"));
}
$this->dbh->commit();
# Fill edges table
$this->dbh->beginTransaction();
$edges = $xp->query("//gexf:graph[1]/gexf:edges/gexf:edge");
foreach($edges as $edge){
$query_insert = "INSERT INTO 'edges' ('importid', 'source', 'target'";
$importid = $this->dbh->quote($edge->getAttribute('id'));
$source = $this->dbh->quote($edge->getAttribute('source'));
$target = $this->dbh->quote($edge->getAttribute('target'));
$query_select = "SELECT $importid, sourcenodes.id, targetnodes.id";
$query_from = " FROM nodes sourcenodes, nodes targetnodes WHERE sourcenodes.importid=$source AND targetnodes.importid=$target";
foreach($edge->getElementsByTagName('attvalue') as $value){
$attName = "attr_".preg_replace('/[^a-z0-9_]*/i','',$value->getAttribute('for'));
$attValue = $this->dbh->quote($value->getAttribute('value'));
$att = $this->getEdgeAttributeByName($attName);
$attName = $this->dbh->quote($att['name']);
if($attName != "''"){
if($att['type'] == 'TAGS'){
$tags = explode("|", $attValue);
foreach($tags as $tag){
$tag = $this->dbh->quote($tag);
$tagQueries[] = "INSERT INTO 'edgestags' (attribute, node, value) SELECT $attId, e.id, $tag FROM edges e WHERE e.importid=$importid";
}
} else {
$query_insert .= ", $attName";
$query_select .= ", $attValue";
}
}
}
$this->dbh->exec(utf8_decode("$query_insert ) $query_select $query_from"));
}
$this->dbh->commit();
# Execute the tag queries
$this->dbh->beginTransaction();
foreach($tagQueries as $query){
$this->dbh->exec(utf8_decode($query));
}
$this->dbh->commit();
} else {
die("unsupported version of GEXF format");
}