SAX parsers all operational

data-collection/BadgesHandler.java

@@ -2,6 +2,7 @@ import org.xml.sax.Attributes;
 import org.xml.sax.SAXException;
 import org.xml.sax.helpers.DefaultHandler;
 
+import javax.xml.XMLConstants;
 import javax.xml.parsers.ParserConfigurationException;
 import javax.xml.parsers.SAXParser;
 import javax.xml.parsers.SAXParserFactory;
@@ -76,7 +77,7 @@ public class BadgesHandler extends DefaultHandler {
                     attr.getValue("Class"),
                     attr.getValue("TagBased")
             );
-            //System.out.println(badgeRow);
+            System.out.println(badgeRow);
 
         } else if (!qName.equals(BADGES)) {
             throw new SAXException(String.format("Unknown tag %s", qName));
@@ -88,6 +89,7 @@ public class BadgesHandler extends DefaultHandler {
         final long start = System.currentTimeMillis();
         System.out.println(start);
         SAXParserFactory factory = SAXParserFactory.newInstance();
+        factory.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING,false);
         SAXParser saxParser = factory.newSAXParser();
         BadgesHandler handler = new BadgesHandler();
         saxParser.parse(

data-collection/CommentsHandler.java

@@ -2,6 +2,7 @@ import org.xml.sax.Attributes;
 import org.xml.sax.SAXException;
 import org.xml.sax.helpers.DefaultHandler;
 
+import javax.xml.XMLConstants;
 import javax.xml.parsers.ParserConfigurationException;
 import javax.xml.parsers.SAXParser;
 import javax.xml.parsers.SAXParserFactory;
@@ -99,6 +100,7 @@ public class CommentsHandler extends DefaultHandler {
         final long start = System.currentTimeMillis();
 
         SAXParserFactory factory = SAXParserFactory.newInstance();
+        factory.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING,false);
         SAXParser saxParser = factory.newSAXParser();
         CommentsHandler handler = new CommentsHandler();
         saxParser.parse(

data-collection/SQLiteSession.java

@@ -33,7 +33,7 @@ public class SQLiteSession {
 
     public static void main(String[] args) {
         SQLiteSession s = new SQLiteSession(
-        "jdbc:sqlite:C:\\Users\\lhb1g20\\OneDrive - University of Southampton\\graph4stackoverflow\\data-collection\\stackoverflow.db"
+        "jdbc:sqlite:..\\stackoverflow.db"
         );
         s.endSession();
     }

data-collection/TagsHandler.java

-public class TagsHandler {
+import org.xml.sax.Attributes;
+import org.xml.sax.SAXException;
+import org.xml.sax.helpers.DefaultHandler;
+
+import javax.xml.XMLConstants;
+import javax.xml.parsers.ParserConfigurationException;
+import javax.xml.parsers.SAXParser;
+import javax.xml.parsers.SAXParserFactory;
+import java.io.IOException;
+
+class TagRow {
+    private String Id;
+    private String TagName;
+    private String Count;
+
+    public TagRow(String Id, String TagName, String Count) {
+        this.Id = Id;
+        this.TagName = TagName;
+        this.Count = Count;
+    }
+
+    public String getId() {
+        return Id;
+    }
+
+    public String getTagName() {
+        return TagName;
+    }
+
+    public String getCount() {
+        return Count;
+    }
+
+    @Override
+    public String toString() {
+        return "TagRow{" +
+                "Id='" + Id + '\'' +
+                ", TagName='" + TagName + '\'' +
+                ", Count='" + Count + '\'' +
+                '}';
+    }
+}
+
+    public class TagsHandler extends DefaultHandler {
+    private static final String ROW = "row";
+    private static final String TAGS = "tags";
+
+    @Override
+    public void startElement(String uri, String lName, String qName, Attributes attr) throws SAXException {
+        if (qName.equals(ROW)) {
+            TagRow badgeRow = new TagRow(
+                    attr.getValue("Id"),
+                    attr.getValue("TagName"),
+                    attr.getValue("Count")
+            );
+            System.out.println(badgeRow);
+
+        } else if (!qName.equals(TAGS)) {
+            throw new SAXException(String.format("Unknown tag %s", qName));
+        }
+
+    }
+
+    public static void main(String[] args) throws ParserConfigurationException, SAXException, IOException {
+        final long start = System.currentTimeMillis();
+        System.out.println(start);
+        SAXParserFactory factory = SAXParserFactory.newInstance();
+        factory.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING,false);
+        SAXParser saxParser = factory.newSAXParser();
+        TagsHandler handler = new TagsHandler();
+        saxParser.parse(
+                "..\\data\\raw\\stackoverflow.com-Tags\\tags.xml", handler);
+
+
+        // time end
+        final long durationInMilliseconds = System.currentTimeMillis()-start;
+        System.out.println("SAX parse took " + durationInMilliseconds + "ms.");
+    }
 }

data-collection/UsersHandler.java

-public class UsersHandler {
+import org.xml.sax.Attributes;
+import org.xml.sax.SAXException;
+import org.xml.sax.helpers.DefaultHandler;
+
+import javax.xml.XMLConstants;
+import javax.xml.parsers.ParserConfigurationException;
+import javax.xml.parsers.SAXParser;
+import javax.xml.parsers.SAXParserFactory;
+import java.io.IOException;
+
+class UserRow {
+    private String Id;
+    private String Reputation;
+    private String CreationDate;
+    private String DisplayName;
+    private String LastAccessDate;
+    private String AboutMe;
+    private String Views;
+    private String UpVotes;
+    private String DownVotes;
+
+
+    public UserRow(String Id, String Reputation, String CreationDate, String DisplayName,
+                   String LastAccessDate, String AboutMe, String Views, String UpVotes,
+                   String DownVotes) {
+        this.Id = Id;
+        this.Reputation = Reputation;
+        this.CreationDate = CreationDate;
+        this.DisplayName = DisplayName;
+        this.LastAccessDate = LastAccessDate;
+        this.AboutMe = AboutMe;
+        this.Views = Views;
+        this.UpVotes = UpVotes;
+        this.DownVotes = DownVotes;
+
+    }
+
+    public String getId() {
+        return Id;
+    }
+
+    public String getReputation() {
+        return Reputation;
+    }
+
+    public String getCreationDate() {
+        return CreationDate;
+    }
+
+    public String getDisplayName() {
+        return DisplayName;
+    }
+
+    public String getLastAccessDate() {
+        return LastAccessDate;
+    }
+
+    public String getAboutMe() {
+        return AboutMe;
+    }
+
+    public String getViews() {
+        return Views;
+    }
+
+    public String getUpVotes() {
+        return UpVotes;
+    }
+
+    public String getDownVotes() {
+        return DownVotes;
+    }
+
+    @Override
+    public String toString() {
+        return "UserRow{" +
+                "Id='" + Id + '\'' +
+                ", Reputation='" + Reputation + '\'' +
+                ", CreationDate='" + CreationDate + '\'' +
+                ", DisplayName='" + DisplayName + '\'' +
+                ", LastAccessDate='" + LastAccessDate + '\'' +
+                ", AboutMe='" + AboutMe + '\'' +
+                ", Views='" + Views + '\'' +
+                ", UpVotes='" + UpVotes + '\'' +
+                ", DownVotes='" + DownVotes + '\'' +
+                '}';
+    }
+}
+
+public class UsersHandler extends DefaultHandler {
+    private static final String ROW = "row";
+    private static final String USERS = "users";
+
+    @Override
+    public void startElement(String uri, String lName, String qName, Attributes attr) throws SAXException {
+        if (qName.equals(ROW)) {
+            UserRow userRow = new UserRow(
+                    attr.getValue("Id"),
+                    attr.getValue("Reputation"),
+                    attr.getValue("CreationDate"),
+                    attr.getValue("DisplayName"),
+                    attr.getValue("LastAccessDate"),
+                    attr.getValue("AboutMe"),
+                    attr.getValue("Views"),
+                    attr.getValue("UpVotes"),
+                    attr.getValue("DownVotes")
+            );
+            System.out.println(userRow);
+
+        } else if (!qName.equals(USERS)) {
+            throw new SAXException(String.format("Unknown tag %s", qName));
+        }
+
+    }
+
+    public static void main(String[] args) throws ParserConfigurationException, SAXException, IOException {
+        final long start = System.currentTimeMillis();
+        System.out.println(start);
+        SAXParserFactory factory = SAXParserFactory.newInstance();
+        factory.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING,false);
+        SAXParser saxParser = factory.newSAXParser();
+        UsersHandler handler = new UsersHandler();
+        saxParser.parse(
+                "..\\data\\raw\\stackoverflow.com-Users\\users.xml", handler);
+
+
+        // time end
+        final long durationInMilliseconds = System.currentTimeMillis()-start;
+        System.out.println("SAX parse took " + durationInMilliseconds + "ms.");
+    }
 }

data-collection/VotesHandler.java

-public class VotesHandler {
+import org.xml.sax.Attributes;
+import org.xml.sax.SAXException;
+import org.xml.sax.helpers.DefaultHandler;
+
+import javax.xml.XMLConstants;
+import javax.xml.parsers.ParserConfigurationException;
+import javax.xml.parsers.SAXParser;
+import javax.xml.parsers.SAXParserFactory;
+import java.io.IOException;
+
+class VoteRow {
+    private String Id;
+    private String PostId;
+    private String VoteTypeId;
+    private String CreationDate;
+
+    public VoteRow(String Id, String PostId, String VoteTypeId, String CreationDate) {
+        this.Id = Id;
+        this.PostId = PostId;
+        this.VoteTypeId = VoteTypeId;
+        this.CreationDate = CreationDate;
+    }
+
+    public String getId() {
+        return Id;
+    }
+
+    public String getPostId() {
+        return PostId;
+    }
+
+    public String getVoteTypeId() {
+        return VoteTypeId;
+    }
+
+    public String getCreationDate() {
+        return CreationDate;
+    }
+
+    @Override
+    public String toString() {
+        return "VoteRow{" +
+                "Id='" + Id + '\'' +
+                ", PostId='" + PostId + '\'' +
+                ", VoteTypeId='" + VoteTypeId + '\'' +
+                ", CreationDate='" + CreationDate + '\'' +
+                '}';
+    }
+}
+
+public class VotesHandler extends DefaultHandler {
+    private static final String ROW = "row";
+    private static final String VOTES = "votes";
+
+    @Override
+    public void startElement(String uri, String lName, String qName, Attributes attr) throws SAXException {
+        if (qName.equals(ROW)) {
+            VoteRow voteRow = new VoteRow(
+                    attr.getValue("Id"),
+                    attr.getValue("PostId"),
+                    attr.getValue("VoteTypeId"),
+                    attr.getValue("CreationDate")
+            );
+            System.out.println(voteRow);
+
+        } else if (!qName.equals(VOTES)) {
+            throw new SAXException(String.format("Unknown tag %s", qName));
+        }
+
+    }
+
+    public static void main(String[] args) throws ParserConfigurationException, SAXException, IOException {
+        final long start = System.currentTimeMillis();
+        System.out.println(start);
+        SAXParserFactory factory = SAXParserFactory.newInstance();
+        factory.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING,false);
+        SAXParser saxParser = factory.newSAXParser();
+        VotesHandler handler = new VotesHandler();
+        saxParser.parse(
+                "..\\data\\raw\\stackoverflow.com-Votes\\votes.xml", handler);
+
+
+        // time end
+        final long durationInMilliseconds = System.currentTimeMillis()-start;
+        System.out.println("SAX parse took " + durationInMilliseconds + "ms.");
+    }
 }