@Override
public UpdateContainer doIt(Workspace workspace) throws CommandException {
Worksheet wk = workspace.getWorksheet(worksheetId);
SuperSelection selection = getSuperSelection(wk);
String Msg =
String.format("begin, Time,%d, Worksheet,%s", System.currentTimeMillis(), worksheetId);
logger.info(Msg);
// Get the HNode
HashMap<String, HashMap<String, String>> rows = new HashMap<String, HashMap<String, String>>();
HNodePath selectedPath = null;
List<HNodePath> columnPaths = wk.getHeaders().getAllPaths();
for (HNodePath path : columnPaths) {
if (path.getLeaf().getId().equals(hNodeId)) {
selectedPath = path;
}
}
// random nodes
Collection<Node> nodes = new ArrayList<Node>();
wk.getDataTable().collectNodes(selectedPath, nodes, selection);
HashSet<Integer> indSet = this.obtainIndexs(nodes.size());
int index = 0;
for (Iterator<Node> iterator = nodes.iterator(); iterator.hasNext(); ) {
Node node = iterator.next();
if (indSet.contains(index)) {
String id = node.getId();
String originalVal = node.getValue().asString();
HashMap<String, String> x = new HashMap<String, String>();
x.put("Org", originalVal);
x.put("Tar", originalVal);
x.put("Orgdis", originalVal);
x.put("Tardis", originalVal);
rows.put(id, x);
}
index++;
}
Msg = String.format("end, Time,%d, Worksheet,%s", System.currentTimeMillis(), worksheetId);
logger.info(Msg);
return new UpdateContainer(new FetchResultUpdate(hNodeId, rows));
}
private void generateRDF(
String wkname,
String query,
List<KR2RMLRDFWriter> writers,
R2RMLMappingIdentifier id,
String baseURI)
throws IOException, JSONException, KarmaException, SQLException, ClassNotFoundException {
logger.debug("Generating RDF...");
WorksheetR2RMLJenaModelParser parserTest = new WorksheetR2RMLJenaModelParser(id);
KR2RMLMapping mapping = parserTest.parse();
AbstractJDBCUtil dbUtil = JDBCUtilFactory.getInstance(dbType);
Connection conn = dbUtil.getConnection(hostname, portnumber, username, password, dBorSIDName);
conn.setAutoCommit(false);
java.sql.Statement stmt =
conn.createStatement(
java.sql.ResultSet.TYPE_FORWARD_ONLY, java.sql.ResultSet.CONCUR_READ_ONLY);
stmt.setFetchSize(DATABASE_TABLE_FETCH_SIZE);
ResultSet r = stmt.executeQuery(query);
ResultSetMetaData meta = r.getMetaData();
;
// Get the column names
List<String> columnNames = new ArrayList<>();
for (int i = 1; i <= meta.getColumnCount(); i++) {
columnNames.add(meta.getColumnName(i));
}
// Prepare required Karma objects
Workspace workspace = initializeWorkspace();
RepFactory factory = workspace.getFactory();
Worksheet wk = factory.createWorksheet(wkname, workspace, encoding);
List<String> headersList = addHeaders(wk, columnNames, factory);
int counter = 0;
ArrayList<String> rowValues = null;
while ((rowValues = dbUtil.parseResultSetRow(r)) != null) {
// Generate RDF and create a new worksheet for every DATABASE_TABLE_FETCH_SIZE rows
if (counter % DATABASE_TABLE_FETCH_SIZE == 0 && counter != 0) {
generateRDFFromWorksheet(wk, workspace, mapping, writers, baseURI);
logger.debug("Done for " + counter + " rows ...");
removeWorkspace(workspace);
parserTest = new WorksheetR2RMLJenaModelParser(id);
mapping = parserTest.parse();
workspace = initializeWorkspace();
factory = workspace.getFactory();
wk = factory.createWorksheet(wkname, workspace, encoding);
headersList = addHeaders(wk, columnNames, factory);
}
/** Add the data * */
Table dataTable = wk.getDataTable();
Row row = dataTable.addRow(factory);
for (int i = 0; i < rowValues.size(); i++) {
row.setValue(headersList.get(i), rowValues.get(i), factory);
}
counter++;
}
generateRDFFromWorksheet(wk, workspace, mapping, writers, baseURI);
// Releasing all the resources
r.close();
conn.close();
stmt.close();
logger.debug("done");
}
@Override
public UpdateContainer doIt(Workspace workspace) throws CommandException {
Worksheet wk = workspace.getWorksheet(worksheetId);
SuperSelection selection = getSuperSelection(wk);
String msg =
String.format(
"Gen rule start,Time,%d, Worksheet,%s", System.currentTimeMillis(), worksheetId);
logger.info(msg);
// Get the HNode
HashMap<String, String> rows = new HashMap<String, String>();
HashMap<String, Integer> amb = new HashMap<String, Integer>();
HNodePath selectedPath = null;
List<HNodePath> columnPaths = wk.getHeaders().getAllPaths();
for (HNodePath path : columnPaths) {
if (path.getLeaf().getId().equals(hNodeId)) {
selectedPath = path;
}
}
Collection<Node> nodes = new ArrayList<Node>();
wk.getDataTable().collectNodes(selectedPath, nodes, selection);
for (Node node : nodes) {
String id = node.getId();
if (!this.nodeIds.contains(id)) continue;
String originalVal = node.getValue().asString();
rows.put(id, originalVal);
this.compResultString += originalVal + "\n";
calAmbScore(id, originalVal, amb);
}
RamblerValueCollection vc = new RamblerValueCollection(rows);
HashMap<String, Vector<String[]>> expFeData = new HashMap<String, Vector<String[]>>();
inputs = new RamblerTransformationInputs(examples, vc);
// generate the program
boolean results = false;
int iterNum = 0;
RamblerTransformationOutput rtf = null;
// initialize the vocabulary
Iterator<String> iterx = inputs.getInputValues().getValues().iterator();
Vector<String> v = new Vector<String>();
int vb_cnt = 0;
while (iterx.hasNext() && vb_cnt < 30) {
String eString = iterx.next();
v.add(eString);
vb_cnt++;
}
Vector<String> vob = UtilTools.buildDict(v);
inputs.setVocab(vob.toArray(new String[vob.size()]));
while (iterNum < 1 && !results) // try to find an program within iterNum
{
rtf = new RamblerTransformationOutput(inputs);
if (rtf.getTransformations().keySet().size() > 0) {
results = true;
}
iterNum++;
}
Iterator<String> iter = rtf.getTransformations().keySet().iterator();
// id:{org: tar: orgdis: tardis: }
HashMap<String, HashMap<String, String>> resdata =
new HashMap<String, HashMap<String, String>>();
HashSet<String> keys = new HashSet<String>();
while (iter.hasNext()) {
String tpid = iter.next();
ValueCollection rvco = rtf.getTransformedValues_debug(tpid);
if (rvco == null) continue;
// constructing displaying data
HashMap<String, String[]> xyzHashMap = new HashMap<String, String[]>();
for (String key : rvco.getNodeIDs()) {
HashMap<String, String> dict = new HashMap<String, String>();
// add to the example selection
boolean isExp = false;
String org = vc.getValue(key);
String classLabel = rvco.getClass(key);
String pretar = rvco.getValue(key);
String dummyValue = pretar;
if (pretar.indexOf("_FATAL_ERROR_") != -1) {
dummyValue = org;
// dummyValue = "#ERROR";
}
try {
UtilTools.StringColorCode(org, dummyValue, dict);
} catch (Exception ex) {
logger.info(String.format("ColorCoding Exception%s, %s", org, dummyValue));
// set dict
dict.put("Org", org);
dict.put("Tar", "ERROR");
dict.put("Orgdis", org);
dict.put("Tardis", "ERROR");
}
for (TransformationExample exp : examples) {
if (exp.getNodeId().compareTo(key) == 0) {
if (!expFeData.containsKey(classLabel)) {
Vector<String[]> vstr = new Vector<String[]>();
String[] texp = {dict.get("Org"), pretar};
vstr.add(texp);
expFeData.put(classLabel, vstr);
} else {
String[] texp = {dict.get("Org"), pretar};
expFeData.get(classLabel).add(texp);
}
isExp = true;
}
}
if (!isExp) {
String[] pair = {dict.get("Org"), dict.get("Tar"), pretar, classLabel};
xyzHashMap.put(key, pair);
}
resdata.put(key, dict);
}
if (!rtf.nullRule) keys.add(getBestExample(xyzHashMap, expFeData));
}
// find the best row
String vars = "";
String expstr = "";
String recmd = "";
for (TransformationExample x : examples) {
expstr += String.format("%s|%s", x.getBefore(), x.getAfter());
}
expstr += "|";
if (rtf.nullRule) {
keys.clear();
// keys.add("-2"); // "-2 indicates null rule"
}
if (!resdata.isEmpty() && !rtf.nullRule) {
recmd = resdata.get(keys.iterator().next()).get("Org");
} else {
recmd = "";
}
msg =
String.format(
"Gen rule end, Time,%d, Worksheet,%s,Examples:%s,Recmd:%s",
System.currentTimeMillis(), worksheetId, expstr, recmd);
logger.info(msg);
return new UpdateContainer(new CleaningResultUpdate(hNodeId, resdata, vars, keys));
}