/**
* The ActionListener implementation
*
* @param event the event.
*/
public void actionPerformed(ActionEvent event) {
String searchText = textField.getText().trim();
if (searchText.equals("") && !saveAs.isSelected() && (fileLength > 10000000)) {
textPane.setText("Blank search text is not allowed for large IdTables.");
} else {
File outputFile = null;
if (saveAs.isSelected()) {
outputFile = chooser.getSelectedFile();
if (outputFile != null) {
String name = outputFile.getName();
int k = name.lastIndexOf(".");
if (k != -1) name = name.substring(0, k);
name += ".txt";
File parent = outputFile.getAbsoluteFile().getParentFile();
outputFile = new File(parent, name);
chooser.setSelectedFile(outputFile);
}
if (chooser.showSaveDialog(this) != JFileChooser.APPROVE_OPTION) System.exit(0);
outputFile = chooser.getSelectedFile();
}
textPane.setText("");
Searcher searcher = new Searcher(searchText, event.getSource().equals(searchPHI), outputFile);
searcher.start();
}
}
// QHandler handles incoming queries.
public void run() {
// hmmm.. this seems like potential bug. I want to check that query is not in table. But even
// if query table contains key,
// that does not necessarily mean it is in table, b/c two queries can have SAME HASHCODE VALUE.
// I need to have some other means.
// Will talk to Rusty @ this on Monday.
if (!qt.containsKey(query)) // check that query is not already in table
{
Searcher.inform(query); // Give information to the Search Monitor panel
NetworkManager.writeButOne(query.getIP(), query); /*Query is forwarded to all connected nodes
except one from which query came. */
qt.put((Packet) query, query); // add query to table, indexed by its unique MessageID
searchResult =
SharedDirectory.search(query.getSearchString()); // check shared directory for query match
numHits = searchResult.getSize();
if (numHits != 0) // package a query hit to send out if there is at least one query match
{
queryID = query.getMessageID();
port = Mine.getPort();
myIP = Mine.getIPAddress();
speed = Mine.getSpeed();
serventID = Mine.getServentIdentifier();
queryHit = new QueryHit(numHits, port, myIP, speed, searchResult, serventID, queryID);
NetworkManager.writeToOne(
query.getIP(), queryHit); // send qHit back to node that sent original query
}
}
}
private Element getDescription(String task) {
if (task != null) {
if (task.equals("define")) return definer.getDescription();
if (task.equals("compare")) return comparer.getDescription();
if (task.equals("search")) return searcher.getDescription();
if (task.equals("wikify")) return wikifier.getDescription();
}
Element description = doc.createElement("Description");
description.appendChild(
createElement(
"Details",
"<p>This servlet provides a range of services for mining information from Wikipedia. Further details depend on what you want to do.</p>"
+ "<p>You can <a href=\""
+ context.getInitParameter("service_name")
+ "?task=search&help\">search for pages</a>, <a href=\""
+ context.getInitParameter("service_name")
+ "?task=compare&help\">measure how terms or articles related to each other</a>, <a href=\""
+ context.getInitParameter("service_name")
+ "?task=define&help\">obtain short definitions from articles</a>, and <a href=\""
+ context.getInitParameter("service_name")
+ "?task=wikify&help\">detect topics in web pages</a>.</p>"));
Element paramTask =
createElement(
"Parameter",
"Specifies what you want to do: can be <em>search</em>, <em>compare</em>, <em>define</em>, or <em>wikify</em>");
paramTask.setAttribute("name", "task");
description.appendChild(paramTask);
Element paramId = doc.createElement("Parameter");
paramId.setAttribute("name", "help");
paramId.appendChild(doc.createTextNode("Specifies that you want help about the service."));
description.appendChild(paramId);
return description;
}
protected void search(boolean persistent, boolean async) {
try {
synchronized (this) {
if (current_search == null) {
current_search = new Searcher(persistent, async);
} else {
if (!current_search.wakeup()) {
current_search = new Searcher(persistent, async);
}
}
}
} catch (Throwable e) {
Debug.out(e);
}
}
public void doGet(HttpServletRequest request, HttpServletResponse response)
throws IOException, ServletException {
try {
response.setHeader("Cache-Control", "no-cache");
response.setCharacterEncoding("UTF-8");
String task = request.getParameter("task");
Element data = null;
// process help request
if (request.getParameter("help") != null) data = getDescription(task);
// redirect to home page if there is no task
if (data == null && task == null) {
response.setContentType("text/html");
response
.getWriter()
.append(
"<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.0 Transitional//EN\"><html><head><meta http-equiv=\"REFRESH\" content=\"0;url="
+ context.getInitParameter("server_path")
+ "></head><body></body></html>");
return;
}
// process definition request
if (data == null && task.equals("define")) {
int id = resolveIntegerArg(request.getParameter("id"), -1);
int length = resolveIntegerArg(request.getParameter("length"), definer.getDefaultLength());
int format = resolveIntegerArg(request.getParameter("format"), definer.getDefaultFormat());
int maxImageWidth =
resolveIntegerArg(
request.getParameter("maxImageWidth"), definer.getDefaultMaxImageWidth());
int maxImageHeight =
resolveIntegerArg(
request.getParameter("maxImageHeight"), definer.getDefaultMaxImageHeight());
int linkDestination =
resolveIntegerArg(
request.getParameter("linkDestination"), definer.getDefaultLinkDestination());
boolean getImages = resolveBooleanArg(request.getParameter("getImages"), false);
data =
definer.getDefinition(
id, length, format, linkDestination, getImages, maxImageWidth, maxImageHeight);
}
// all of the remaining tasks require data to be cached, so lets make sure that is finished
// before continuing.
if (!cachingThread.isOk()) throw new ServletException("Could not cache wikipedia data");
double progress = cachingThread.getProgress();
if (data == null && (progress < 1 || task.equals("progress"))) {
// still caching up data, not ready to return a response yet.
data = doc.createElement("loading");
data.setAttribute("progress", df.format(progress));
task = "loading";
}
// process search request
if (data == null && task.equals("search")) {
String term = request.getParameter("term");
String id = request.getParameter("id");
int linkLimit =
resolveIntegerArg(request.getParameter("linkLimit"), searcher.getDefaultMaxLinkCount());
int senseLimit =
resolveIntegerArg(
request.getParameter("senseLimit"), searcher.getDefaultMaxSenseCount());
if (id == null) data = searcher.doSearch(term, linkLimit, senseLimit);
else data = searcher.doSearch(Integer.parseInt(id), linkLimit);
}
// process compare request
if (data == null && task.equals("compare")) {
String term1 = request.getParameter("term1");
String term2 = request.getParameter("term2");
int linkLimit =
resolveIntegerArg(request.getParameter("linkLimit"), comparer.getDefaultMaxLinkCount());
boolean details =
resolveBooleanArg(request.getParameter("details"), comparer.getDefaultShowDetails());
data = comparer.getRelatedness(term1, term2, details, linkLimit);
}
// process wikify request
if (data == null && task.equals("wikify")) {
if (this.wikifier == null)
throw new ServletException(
"Wikifier is not available. You must configure the servlet so that it has access to link detection and disambiguation models.");
String source = request.getParameter("source");
int sourceMode =
resolveIntegerArg(request.getParameter("sourceMode"), Wikifier.SOURCE_AUTODETECT);
String linkColor = request.getParameter("linkColor");
String baseColor = request.getParameter("baseColor");
double minProb =
resolveDoubleArg(
request.getParameter("minProbability"), wikifier.getDefaultMinProbability());
int repeatMode =
resolveIntegerArg(request.getParameter("repeatMode"), wikifier.getDefaultRepeatMode());
boolean showTooltips =
resolveBooleanArg(
request.getParameter("showTooltips"), wikifier.getDefaultShowTooltips());
String bannedTopics = request.getParameter("bannedTopics");
boolean wrapInXml = resolveBooleanArg(request.getParameter("wrapInXml"), true);
if (wrapInXml) {
data =
wikifier.wikifyAndWrapInXML(
source,
sourceMode,
minProb,
repeatMode,
bannedTopics,
baseColor,
linkColor,
showTooltips);
} else {
response.setContentType("text/html");
response
.getWriter()
.append(
wikifier.wikify(
source,
sourceMode,
minProb,
repeatMode,
bannedTopics,
baseColor,
linkColor,
showTooltips));
return;
}
}
if (data == null) throw new Exception("Unknown Task");
// wrap data
Element wrapper = doc.createElement("WikipediaMinerResponse");
wrapper.setAttribute("server_path", context.getInitParameter("server_path"));
wrapper.setAttribute("service_name", context.getInitParameter("service_name"));
wrapper.appendChild(data);
data = wrapper;
// Transform or serialize xml data as appropriate
Transformer tf = null;
if (request.getParameter("xml") == null) {
// we need to transform the data into html
tf = transformersByName.get(task);
if (request.getParameter("help") != null) tf = transformersByName.get("help");
}
if (tf == null) {
// we need to serialize the data as xml
tf = transformersByName.get("serializer");
response.setContentType("application/xml");
} else {
// output will be transformed to html
response.setContentType("text/html");
response
.getWriter()
.append(
"<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.01//EN\" \"http://www.w3.org/TR/html4/strict.dtd\">\n");
}
tf.transform(new DOMSource(data), new StreamResult(response.getWriter()));
} catch (Exception error) {
response.reset();
response.setContentType("application/xml");
response.setHeader("Cache-Control", "no-cache");
response.setCharacterEncoding("UTF8");
Element xmlError = doc.createElement("Error");
if (error.getMessage() != null) xmlError.setAttribute("message", error.getMessage());
Element xmlStackTrace = doc.createElement("StackTrace");
xmlError.appendChild(xmlStackTrace);
for (StackTraceElement ste : error.getStackTrace()) {
Element xmlSte = doc.createElement("StackTraceElement");
xmlSte.setAttribute("message", ste.toString());
xmlStackTrace.appendChild(xmlSte);
}
try {
transformersByName
.get("serializer")
.transform(new DOMSource(xmlError), new StreamResult(response.getWriter()));
} catch (Exception e) {
// TODO: something for when an error is thrown processing an error????
}
;
}
}
// private static int[] oldToNew(IndexReader reader, Searcher searcher) throws IOException {
private static DocScore[] newToOld(IndexReader reader, Searcher searcher) throws IOException {
int readerMax = reader.maxDoc();
DocScore[] newToOld = new DocScore[readerMax];
// use site, an indexed, un-tokenized field to get boost
// byte[] boosts = reader.norms("site"); TODO MC
/* TODO MC */
Document docMeta;
Pattern includes = Pattern.compile("\\|");
String value = NutchConfiguration.create().get(INCLUDE_EXTENSIONS_KEY, "");
String includeExtensions[] = includes.split(value);
Hashtable<String, Boolean> validExtensions = new Hashtable<String, Boolean>();
for (int i = 0; i < includeExtensions.length; i++) {
validExtensions.put(includeExtensions[i], true);
System.out.println("extension boosted " + includeExtensions[i]);
}
/* TODO MC */
for (int oldDoc = 0; oldDoc < readerMax; oldDoc++) {
float score;
if (reader.isDeleted(oldDoc)) {
// score = 0.0f;
score = -1f; // TODO MC
} else {
// score = Similarity.decodeNorm(boosts[oldDoc]); TODO MC
/* TODO MC */
docMeta = searcher.doc(oldDoc);
if (validExtensions.get(docMeta.get("subType"))
== null) { // searched extensions will have higher scores
score = -0.5f;
} else {
score = Integer.parseInt(docMeta.get("inlinks"));
/*
if (score==0) {
score=0.001f; // TODO MC - to not erase
}
*/
}
/* TODO MC */
// System.out.println("Score for old document "+oldDoc+" is "+score+" and type
// "+docMeta.get("subType")); // TODO MC debug remove
}
DocScore docScore = new DocScore();
docScore.doc = oldDoc;
docScore.score = score;
newToOld[oldDoc] = docScore;
}
System.out.println("Sorting " + newToOld.length + " documents.");
Arrays.sort(newToOld);
// HeapSorter.sort(newToOld); // TODO MC - due to the lack of space
/* TODO MC
int[] oldToNew = new int[readerMax];
for (int newDoc = 0; newDoc < readerMax; newDoc++) {
DocScore docScore = newToOld[newDoc];
//oldToNew[docScore.oldDoc] = docScore.score > 0.0f ? newDoc : -1; // TODO MC
oldToNew[docScore.oldDoc] = newDoc; // TODO MC
}
*/
/* TODO MC *
for (int newDoc = 0; newDoc < readerMax; newDoc++) {
DocScore docScore = newToOld[newDoc];
System.out.println("Score for new document "+newDoc+" is "+docScore.score); // TODO MC debug remove
}
* TODO MC */
// return oldToNew; TODO MC
return newToOld; // TODO MC
}