diff --git a/.gitignore b/.gitignore
index 723ef36f4e4f32c4560383aa5987c575a30c6535..1a192ea1a341a08aae7411cf09df2d2bb0b63dc4 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1 +1,3 @@
-.idea
\ No newline at end of file
+.idea
+data/
+data-collection/out/
\ No newline at end of file
diff --git a/data-collection/BadgesHandler.java b/data-collection/BadgesHandler.java
new file mode 100644
index 0000000000000000000000000000000000000000..0074ce4010d6c058b2e63e51968bb8ab9fb6e9b4
--- /dev/null
+++ b/data-collection/BadgesHandler.java
@@ -0,0 +1,101 @@
+import org.xml.sax.Attributes;
+import org.xml.sax.SAXException;
+import org.xml.sax.helpers.DefaultHandler;
+
+import javax.xml.parsers.ParserConfigurationException;
+import javax.xml.parsers.SAXParser;
+import javax.xml.parsers.SAXParserFactory;
+import java.io.IOException;
+
+class BadgeRow {
+ private String Id;
+ private String UserId;
+ private String Name;
+ private String Date;
+ private String BadgeClass;
+ private String TagBased;
+
+ public BadgeRow(String id, String userId, String name, String date, String badgeClass, String tagBased) {
+ Id = id;
+ UserId = userId;
+ Name = name;
+ Date = date;
+ BadgeClass = badgeClass;
+ TagBased = tagBased;
+ }
+
+ public String getId() {
+ return Id;
+ }
+
+ public String getUserId() {
+ return UserId;
+ }
+
+ public String getName() {
+ return Name;
+ }
+
+ public String getDate() {
+ return Date;
+ }
+
+ public String getBadgeClass() {
+ return BadgeClass;
+ }
+
+ public String getTagBased() {
+ return TagBased;
+ }
+
+ @Override
+ public String toString() {
+ return "BadgeRow{" +
+ "Id='" + Id + '\'' +
+ ", UserId='" + UserId + '\'' +
+ ", Name='" + Name + '\'' +
+ ", Date='" + Date + '\'' +
+ ", BadgeClass='" + BadgeClass + '\'' +
+ ", TagBased='" + TagBased + '\'' +
+ '}';
+ }
+}
+
+public class BadgesHandler extends DefaultHandler {
+ private static final String ROW = "row";
+ private static final String BADGES = "badges";
+
+ @Override
+ public void startElement(String uri, String lName, String qName, Attributes attr) throws SAXException {
+ if (qName.equals(ROW)) {
+ BadgeRow badgeRow = new BadgeRow(
+ attr.getValue("Id"),
+ attr.getValue("UserId"),
+ attr.getValue("Name"),
+ attr.getValue("Date"),
+ attr.getValue("Class"),
+ attr.getValue("TagBased")
+ );
+ //System.out.println(badgeRow);
+
+ } else if (!qName.equals(BADGES)) {
+ 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();
+ SAXParser saxParser = factory.newSAXParser();
+ BadgesHandler handler = new BadgesHandler();
+ saxParser.parse(
+ "..\\data\\raw\\stackoverflow.com-Badges\\badges.xml", handler);
+
+
+ // time end
+ final long durationInMilliseconds = System.currentTimeMillis()-start;
+ System.out.println("SAX parse took " + durationInMilliseconds + "ms.");
+ }
+}
diff --git a/data-collection/CommentsHandler.java b/data-collection/CommentsHandler.java
new file mode 100644
index 0000000000000000000000000000000000000000..6c038ee4cd6f2ce9f6526e6a452d4a7f6e860782
--- /dev/null
+++ b/data-collection/CommentsHandler.java
@@ -0,0 +1,112 @@
+import org.xml.sax.Attributes;
+import org.xml.sax.SAXException;
+import org.xml.sax.helpers.DefaultHandler;
+
+import javax.xml.parsers.ParserConfigurationException;
+import javax.xml.parsers.SAXParser;
+import javax.xml.parsers.SAXParserFactory;
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.List;
+
+class CommentRow {
+ private String id;
+ private String postId;
+ private String score;
+ private String text;
+ private String creationDate;
+ private String userId;
+ private String contentLicense;
+
+ public CommentRow(String id, String postId, String score, String text,
+ String creationDate, String userId, String contentLicense) {
+ this.id = id;
+ this.postId = postId;
+ this.score = score;
+ this.text = text;
+ this.creationDate = creationDate;
+ this.userId = userId;
+ this.contentLicense = contentLicense;
+ }
+
+ public String getId() {
+ return id;
+ }
+
+ public String getPostId() {
+ return postId;
+ }
+
+ public String getScore() {
+ return score;
+ }
+
+ public String getText() {
+ return text;
+ }
+
+ public String getCreationDate() {
+ return creationDate;
+ }
+
+ public String getUserId() {
+ return userId;
+ }
+
+ public String getContentLicense() {
+ return contentLicense;
+ }
+
+ @Override
+ public String toString() {
+ return "CommentRow{" +
+ "id='" + id + '\'' +
+ ", postId='" + postId + '\'' +
+ ", score='" + score + '\'' +
+ ", text='" + text + '\'' +
+ ", creationDate='" + creationDate + '\'' +
+ ", userId='" + userId + '\'' +
+ ", contentLicense='" + contentLicense + '\'' +
+ '}';
+ }
+}
+public class CommentsHandler extends DefaultHandler {
+ private static final String ROW = "row";
+ private static final String COMMENTS = "comments";
+ private StringBuilder elementValue;
+
+ @Override
+ public void startElement(String uri, String lName, String qName, Attributes attr) throws SAXException {
+ if (qName.equals(ROW)) {
+ CommentRow commentRow = new CommentRow(
+ attr.getValue("Id"),
+ attr.getValue("PostId"),
+ attr.getValue("Score"),
+ attr.getValue("Text"),
+ attr.getValue("CreationDate"),
+ attr.getValue("UserId"),
+ attr.getValue("ContentLicense")
+ );
+ System.out.println(commentRow);
+
+ } else if (!qName.equals(COMMENTS)) {
+ throw new SAXException(String.format("Unknown tag %s", qName));
+ }
+
+ }
+
+ public static void main(String[] args) throws ParserConfigurationException, SAXException, IOException {
+ final long start = System.currentTimeMillis();
+
+ SAXParserFactory factory = SAXParserFactory.newInstance();
+ SAXParser saxParser = factory.newSAXParser();
+ CommentsHandler handler = new CommentsHandler();
+ saxParser.parse(
+ "..\\data\\raw\\stackoverflow.com-Comments\\comments.xml", handler);
+
+
+ // time end
+ final long durationInMilliseconds = System.currentTimeMillis()-start;
+ System.out.println("SAX parse took " + durationInMilliseconds + "ms.");
+ }
+}
diff --git a/data-collection/PostsHandler.java b/data-collection/PostsHandler.java
new file mode 100644
index 0000000000000000000000000000000000000000..47b7063534e7f845771c02577b0030d58d03007f
--- /dev/null
+++ b/data-collection/PostsHandler.java
@@ -0,0 +1,206 @@
+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 PostRow {
+ private String Id;
+ private String PostTypeId;
+ private String AcceptedAnswerId;
+ private String CreationDate;
+ private String Score;
+ private String ViewCount;
+ private String Body;
+ private String OwnerUserId;
+ private String LastEditorUserId;
+ private String LastEditorDisplayName;
+ private String LastEditDate;
+ private String Title;
+ private String Tags;
+ private String AnswerCount;
+ private String CommentCount;
+ private String FavoriteCount;
+ private String CommunityOwnedDate;
+ private String ContentLicense;
+
+ public PostRow(String id, String postTypeId, String acceptedAnswerId,
+ String creationDate, String score, String viewCount,
+ String body, String ownerUserId, String lastEditorUserId,
+ String lastEditorDisplayName, String lastEditDate,
+ String title, String tags, String answerCount,
+ String commentCount, String favoriteCount,
+ String communityOwnedDate, String contentLicense) {
+ Id = id;
+ PostTypeId = postTypeId;
+ AcceptedAnswerId = acceptedAnswerId;
+ CreationDate = creationDate;
+ Score = score;
+ ViewCount = viewCount;
+ Body = body;
+ OwnerUserId = ownerUserId;
+ LastEditorUserId = lastEditorUserId;
+ LastEditorDisplayName = lastEditorDisplayName;
+ LastEditDate = lastEditDate;
+ Title = title;
+ Tags = tags;
+ AnswerCount = answerCount;
+ CommentCount = commentCount;
+ FavoriteCount = favoriteCount;
+ CommunityOwnedDate = communityOwnedDate;
+ ContentLicense = contentLicense;
+ }
+
+ @Override
+ public String toString() {
+ return "PostRow{" +
+ "Id='" + Id + '\'' +
+ ", PostTypeId='" + PostTypeId + '\'' +
+ ", AcceptedAnswerId='" + AcceptedAnswerId + '\'' +
+ ", CreationDate='" + CreationDate + '\'' +
+ ", Score='" + Score + '\'' +
+ ", ViewCount='" + ViewCount + '\'' +
+ ", Body='" + Body + '\'' +
+ ", OwnerUserId='" + OwnerUserId + '\'' +
+ ", LastEditorUserId='" + LastEditorUserId + '\'' +
+ ", LastEditorDisplayName='" + LastEditorDisplayName + '\'' +
+ ", LastEditDate='" + LastEditDate + '\'' +
+ ", Title='" + Title + '\'' +
+ ", Tags='" + Tags + '\'' +
+ ", AnswerCount='" + AnswerCount + '\'' +
+ ", CommentCount='" + CommentCount + '\'' +
+ ", FavoriteCount='" + FavoriteCount + '\'' +
+ ", CommunityOwnedDate='" + CommunityOwnedDate + '\'' +
+ ", ContentLicense='" + ContentLicense + '\'' +
+ '}';
+ }
+
+ public String getId() {
+ return Id;
+ }
+
+ public String getPostTypeId() {
+ return PostTypeId;
+ }
+
+ public String getAcceptedAnswerId() {
+ return AcceptedAnswerId;
+ }
+
+ public String getCreationDate() {
+ return CreationDate;
+ }
+
+ public String getScore() {
+ return Score;
+ }
+
+ public String getViewCount() {
+ return ViewCount;
+ }
+
+ public String getBody() {
+ return Body;
+ }
+
+ public String getOwnerUserId() {
+ return OwnerUserId;
+ }
+
+ public String getLastEditorUserId() {
+ return LastEditorUserId;
+ }
+
+ public String getLastEditorDisplayName() {
+ return LastEditorDisplayName;
+ }
+
+ public String getLastEditDate() {
+ return LastEditDate;
+ }
+
+ public String getTitle() {
+ return Title;
+ }
+
+ public String getTags() {
+ return Tags;
+ }
+
+ public String getAnswerCount() {
+ return AnswerCount;
+ }
+
+ public String getCommentCount() {
+ return CommentCount;
+ }
+
+ public String getFavoriteCount() {
+ return FavoriteCount;
+ }
+
+ public String getCommunityOwnedDate() {
+ return CommunityOwnedDate;
+ }
+
+ public String getContentLicense() {
+ return ContentLicense;
+ }
+}
+
+public class PostsHandler extends DefaultHandler {
+ private static final String ROW = "row";
+ private static final String POSTS = "posts";
+ private StringBuilder elementValue;
+
+ @Override
+ public void startElement(String uri, String lName, String qName, Attributes attr) throws SAXException {
+ if (qName.equals(ROW)) {
+ PostRow postRow = new PostRow(
+ attr.getValue("Id"),
+ attr.getValue("PostTypeId"),
+ attr.getValue("AcceptedAnswerId"),
+ attr.getValue("CreationDate"),
+ attr.getValue("Score"),
+ attr.getValue("ViewCount"),
+ attr.getValue("Body"),
+ attr.getValue("OwnerUserId"),
+ attr.getValue("LastEditorUserId"),
+ attr.getValue("LastEditorDisplayName"),
+ attr.getValue("LastEditDate"),
+ attr.getValue("Title"),
+ attr.getValue("Tags"),
+ attr.getValue("AnswerCount"),
+ attr.getValue("CommentCount"),
+ attr.getValue("FavoriteCount"),
+ attr.getValue("CommunityOwnedDate"),
+ attr.getValue("ContentLicense")
+ );
+ System.out.println(postRow);
+
+ } else if (!qName.equals(POSTS)) {
+ throw new SAXException(String.format("Unknown tag %s", qName));
+ }
+
+ }
+
+ public static void main(String[] args) throws ParserConfigurationException, SAXException, IOException {
+ final long start = System.currentTimeMillis();
+
+ SAXParserFactory factory = SAXParserFactory.newInstance();
+ factory.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING,false);
+ SAXParser saxParser = factory.newSAXParser();
+ PostsHandler handler = new PostsHandler();
+ saxParser.parse(
+ "..\\data\\raw\\stackoverflow.com-Posts\\posts.xml", handler);
+
+
+ // time end
+ final long durationInMilliseconds = System.currentTimeMillis()-start;
+ System.out.println("SAX parse took " + durationInMilliseconds + "ms.");
+ }
+}
diff --git a/data-collection/SQLiteSession.java b/data-collection/SQLiteSession.java
new file mode 100644
index 0000000000000000000000000000000000000000..313845e45c197ce0b8e2f471fa17e9628adeb6f6
--- /dev/null
+++ b/data-collection/SQLiteSession.java
@@ -0,0 +1,40 @@
+import org.apache.logging.log4j.Level;
+import org.apache.logging.log4j.LogManager;
+import org.apache.logging.log4j.Logger;
+import org.apache.logging.log4j.core.config.Configurator;
+
+import java.sql.Connection;
+import java.sql.DriverManager;
+import java.sql.SQLException;
+
+public class SQLiteSession {
+ private final Logger logger = LogManager.getLogger(SQLiteSession.class);
+ private Connection conn;
+ public SQLiteSession(String url) {
+ Configurator.setLevel(logger.getName(), Level.INFO);
+ try {
+ // create a connection to the database
+ conn = DriverManager.getConnection(url);
+ logger.info("Connection to SQLite has been established.");
+ } catch (SQLException e) {
+ logger.error(e.getMessage());
+ }
+ }
+
+ public void endSession() {
+ try {
+ if (conn != null) {
+ conn.close();
+ }
+ } catch (SQLException ex) {
+ logger.error(ex.getMessage());
+ }
+ }
+
+ public static void main(String[] args) {
+ SQLiteSession s = new SQLiteSession(
+ "jdbc:sqlite:C:\\Users\\lhb1g20\\OneDrive - University of Southampton\\graph4stackoverflow\\data-collection\\stackoverflow.db"
+ );
+ s.endSession();
+ }
+}
\ No newline at end of file
diff --git a/data-collection/TagsHandler.java b/data-collection/TagsHandler.java
new file mode 100644
index 0000000000000000000000000000000000000000..02188aee3c67ec56ecaecb1bd073d6e41ad2ba9b
--- /dev/null
+++ b/data-collection/TagsHandler.java
@@ -0,0 +1,2 @@
+public class TagsHandler {
+}
diff --git a/data-collection/UsersHandler.java b/data-collection/UsersHandler.java
new file mode 100644
index 0000000000000000000000000000000000000000..2714e4b21bc388806be3b1104d12af14becd56f7
--- /dev/null
+++ b/data-collection/UsersHandler.java
@@ -0,0 +1,2 @@
+public class UsersHandler {
+}
diff --git a/data-collection/VotesHandler.java b/data-collection/VotesHandler.java
new file mode 100644
index 0000000000000000000000000000000000000000..89a990d8762b81ff13503166fede911fc4a4bf4f
--- /dev/null
+++ b/data-collection/VotesHandler.java
@@ -0,0 +1,2 @@
+public class VotesHandler {
+}
diff --git a/data-collection/sqlite-jdbc-3.39.3.0.jar b/data-collection/sqlite-jdbc-3.39.3.0.jar
new file mode 100644
index 0000000000000000000000000000000000000000..23ab07358494cdcad77722f4b08748c1442b512a
Binary files /dev/null and b/data-collection/sqlite-jdbc-3.39.3.0.jar differ
diff --git a/data-collection/stackoverflow.db b/data-collection/stackoverflow.db
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391