--- /dev/null
+/**
+ * A FilterResult encapsulates the result of a filter made by checking a mail.
+ **/
+public class FilterResult {
+ /**
+ * This value is used if type is ERROR or UNKNOWN.
+ */
+ public double NO_RESULT;
+
+ /**
+ * A result value greater or equal this value indicates that the filter has
+ * decided on spam.
+ */
+ public double SPAM_THRESHOLD;
+ public double ABSOLUTE_SPAM;
+ public double ABSOLUTE_HAM;
+
+ //TODO decide a good way of deciding
+ public double result; // the result, a value between 0 (ham) and 1 (spam), negative values for "error", "unknown" etc.
+
+ //public HashMap<String,String> properties = new HashMap<String,String>(); // additional properties of the filter (mainly for statistics)
+
+ // -----------------------------------------------------------------------------
+
+ public FilterResult(double result) {
+ SPAM_THRESHOLD=0.5;
+ ABSOLUTE_SPAM=1.0;
+ ABSOLUTE_HAM=0.0;
+ NO_RESULT=-1;
+ this.result = result;
+ }
+
+ public double getResult() {
+ return result;
+ }
+
+ public boolean isSpam() {
+ return result >= SPAM_THRESHOLD;
+ }
+
+ /*
+ public void addProperty(String key, String value) {
+ properties.put(key,value);
+ }
+
+ public String getProperty(String key) {
+ return properties.get(key);
+ }
+
+ public HashMap<String,String> getProperties() {
+ return properties;
+ }
+ */
+}
--- /dev/null
+/**
+ * This class is a container for all data contained in an Email Message.
+ **/
+public class Mail {
+
+ String header; // the full header
+ //String sentOn; // time the message was sent
+ //String receivedOn; // time when the message arrived
+ String from; // the "from" field
+ String to; // the "to" field
+ String cc;
+ String subject;
+ String body;
+ String sourceCode;
+ boolean hasAttachement;
+ //String encoding; //rich text, plain, html
+
+ String messageID; // cached message ID for reuse (takes a lot of memory and is used all over the place)
+ //same as hashcode of a class
+
+ public Mail() {
+ messageID=null;
+ }
+
+ // -------------------------------------------------------
+
+ public void setHeader(String header) {
+ this.header = header;
+ }
+
+ public String getHeader() {
+ return header;
+ }
+
+ public void setSentOn(String sentOn) {
+ this.sentOn = sentOn;
+ }
+
+ public String getSentOn() {
+ return sentOn;
+ }
+
+ /*
+ public Date getSentOnAsDate() {
+ String sentOn = getSentOn();
+ return parseDate(sentOn);
+ }
+
+ public void setReceivedOn(String receivedOn) {
+ this.receivedOn = receivedOn;
+ }
+
+ public String getReceivedOn() {
+ return receivedOn;
+ }
+
+ public Date getReceivedOnAsDate() {
+ String receivedOn = getReceivedOn();
+ return parseDate(receivedOn);
+ }
+ */
+
+ /**
+ * Parses a given Date-String in into a real Date-Object
+ *
+ * @param stringDate the string in format dd.mm.yyyy hh:mm
+ * @return a Date containing the info of the string or the actual date and time if something fails.
+ */
+ /*
+ public Date parseDate(String stringDate) {
+ // date is in this format: dd.mm.yyyy hh:mm
+ if (stringDate == null || "N/A".equals(stringDate)) {
+ return new Date();
+ }
+ try {
+ synchronized (MAIL_TIME_FORMAT) {
+ return MAIL_TIME_FORMAT.parse(stringDate);
+ }
+ } catch (Throwable e) {
+ return new Date();
+ }
+ }
+ */
+
+ public void setFrom(String from) {
+ this.from = from;
+ }
+
+ public String getFrom() {
+ return from;
+ }
+
+ public void setTo(String to) {
+ this.to = to;
+ }
+
+ public String getTo() {
+ return to;
+ }
+
+ public void setCc(String cc) {
+ this.cc = cc;
+ }
+
+ public String getCc() {
+ return cc;
+ }
+
+ public void setSubject(String subject) {
+ this.subject = subject;
+ }
+
+ public String getSubject() {
+ return subject;
+ }
+
+ public void setBody(String body) {
+ this.body = body;
+ }
+
+ public String getBody() {
+ return body;
+ }
+
+ public void setSourceCode(String sourceCode) {
+ this.sourceCode = sourceCode;
+ }
+
+ public String getSourceCode() {
+ return sourceCode;
+ }
+
+ // TODO: String? Is this a boolean, a number, or can be both?
+ public void setHasAttachement(boolean hasAttachement) {
+ this.hasAttachement = hasAttachement;
+ }
+
+ public boolean getHasAttachement() {
+ return hasAttachement;
+ }
+
+ public void setEncoding(String encoding) {
+ this.encoding = encoding;
+ }
+
+ public String getEncoding() {
+ return encoding;
+ }
+
+ public boolean isTextEncoding() {
+ return getEncoding().toLowerCase().indexOf("plain") >= 0;
+ }
+
+ public boolean isHTMLEncoding() {
+ return getEncoding().toLowerCase().indexOf("html") >= 0;
+ }
+
+ public String toString() {
+ return getBody() + "," + getCc() + "," + getEncoding() + "," + getFrom() + "," + getHasAttachement() + "," + getHeader() + "," + getReceivedOn() + "," + getSentOn() + "," + getSourceCode() + "," + getSubject() + "," + getTo();
+ }
+
+ /*
+ public String getID() {
+ if (messageID == null) { // no cached version
+ // Take the message-ID header as ID (if present)
+ String[] messageIDs = getHeaderField("Message-ID");
+ if ((messageIDs != null) && (messageIDs.length > 0)) {
+ messageID = messageIDs[0];
+ } else { // otherwise, hash header and body as ID
+ return String.valueOf(getHeader().hashCode() + getBody().hashCode());
+ }
+ }
+
+ return messageID;
+ }
+ */
+
+ public String[] getHeaderField(String fieldName) {
+
+ }
+
+ public String extractEMailAddress() {
+
+ }
+
+ public boolean equals(Object o) {
+ if (o instanceof Mail) {
+ Mail mail = (Mail)o;
+ return this.getID().equals(mail.getID());
+ }
+
+ return false;
+ }
+}
--- /dev/null
+public class SpamFilter extends Thread {
+ DistributedHashMap mydhmap;
+ int id; //thread id
+ int numiter;
+ int numemail;
+ /**
+ * Total number of threads
+ **/
+ int nthreads;
+
+ public SpamFilter() {
+
+ }
+
+ public SpamFilter(int numiter, int numemail,int threadid) {
+ this.numiter=numiter;
+ this.numemail=numemail;
+ this.id = id;
+ }
+
+ public void run() {
+ int niter;
+ int nemails
+ atomic {
+ niter=numiter;
+ nemails=numemails;
+ }
+
+ Random rand = new Random(0);
+
+ for(int i=0; i<niter; i++) {
+ for(int j=0; j<nemails; j++) {
+ int pickemail = rand.nextInt(100);
+ //String email = getEmail(pickemail);
+ //checkMails(email);
+ }
+ }
+ }
+
+ public static void main(String[] args) {
+ int nthreads;
+ int[] mid = new int[8];
+ mid[0] = (128<<24)|(195<<16)|(136<<8)|162; //dc-1.calit2
+ mid[1] = (128<<24)|(195<<16)|(136<<8)|163; //dc-2.calit2
+ mid[2] = (128<<24)|(195<<16)|(136<<8)|164; //dc-3.calit2
+ mid[3] = (128<<24)|(195<<16)|(136<<8)|165; //dc-4.calit2
+ mid[4] = (128<<24)|(195<<16)|(136<<8)|166; //dc-5.calit2
+ mid[5] = (128<<24)|(195<<16)|(136<<8)|167; //dc-6.calit2
+ mid[6] = (128<<24)|(195<<16)|(136<<8)|168; //dc-7.calit2
+ mid[7] = (128<<24)|(195<<16)|(136<<8)|169; //dc-8.calit2
+
+
+ /**
+ * Read options from command prompt
+ **/
+ SpamFilter sf = new SpamFilter();
+ SpamFilter.parseCmdLine(args, sf);
+
+ /**
+ * Create Global data structure
+ **/
+ DistributedHashMap dhmap;
+ atomic {
+ dhmap = global new DistributedHashMap(500, 0.75f);
+ }
+ //3. N times iteration of work that needs to be done
+ // by each client
+
+ }
+
+ public static void parseCmdLine(String args[], SpamFilter sf) {
+ int i = 0;
+ String arg;
+ while (i < args.length && args[i].startsWith("-")) {
+ arg = args[i++];
+ //check options
+ if(arg.equals("-n")) { //num of iterations
+ if(i < args.length) {
+ sf.numiter = new Integer(args[i++]).intValue();
+ }
+ } else if(arg.equals("-e")) { //num of emails
+ if(i < args.length) {
+ sf.numemail = new Integer(args[i++]).intValue();
+ }
+ } else if(arg.equals("-t")) { //num of threads
+ if(i < args.length) {
+ sf.threshold = new Integer(args[i++]).intValue();
+ }
+ } else if(arg.equals("-h")) {
+ sf.usage();
+ }
+ }
+ if(sf.nthreads == 0) {
+ sf.usage();
+ }
+ }
+
+ /**
+ * The usage routine describing the program
+ **/
+ public void usage() {
+ System.out.println("usage: ./spamfilter -n <num iterations> -e <num emails> -t <num threads>\n");
+ System.out.println( " -n : num iterations");
+ System.out.println( " -e : number of emails");
+ System.out.println( " -t : number of threads");
+ }
+
+ /**
+ * Returns signatures to the Spam filter
+ **/
+ public FilterResult[] checkMail(Mail mail) {
+ //Preprocess emails
+ //StringBuffer[] mailStrings = createMailStrings();
+ //Compute signatures
+ //CommunicationEngine checkEngine = getCheckEngine();
+ //SignatureComputer sigComp = new SignatureComputer();
+ //check with global data structure
+
+ //return results
+ FilterResult[] filterResults = new FilterResult[mailStrings.length];
+
+ return filterResults;
+ }
+}
projects / IRC.git / commitdiff