updateTokenUsingCrawledData (82a7c7cc) · Commits · EC504 Spring 2024 Group Projects / Group6

CheckerCorrector/Checker.java

+19 −1

Original line number	Diff line number	Diff line
		@@ -21,7 +21,25 @@ public class Checker {
		DBinterface dbInterface = new DBinterface();
		DirectedGraph<State> graph = basicGraphClass.getGraph();
		JsonMaker jsonMaker = JsonMaker.create();
		if(argPars.isUpdateHashTable()){
		if(argPars.isUpdateToken()){
		if(argPars.isCheckFile()){
		SentenceExtractor extractor = SentenceExtractor.of(argPars.getFileName());
		List<String> extractedSentences = extractor.getSentences();
		for (String sentence : extractedSentences) {
		dbInterface.updateTokenInDatabase(sentence.toLowerCase(), graph);
		PhraseExtractor extractorPhrase = PhraseExtractor.fromSentence(sentence, 3, 5);
		List<String> phrases = extractorPhrase.getPhrases();
		for (String phrase : phrases) {
		dbInterface.updateTokenInDatabase(phrase.toLowerCase(), graph);
		}
		}
		}else if(argPars.isCheckSentence()){
		dbInterface.updateTokenInDatabase(argPars.getSentence().toLowerCase(), graph);
		for (String phrase : PhraseExtractor.fromSentence(argPars.getSentence(),3, 5).getPhrases()) {
		//dbInterface.updateTokenInDatabase(phrase.toLowerCase(), graph);
		}
		}
		}else if(argPars.isUpdateHashTable()){
		if(argPars.isCheckFile()){
		SentenceExtractor extractor = SentenceExtractor.of(argPars.getFileName());
		List<String> extractedSentences = extractor.getSentences();

CheckerCorrector/DBinterface/DBinterface.java

+117 −0

Original line number	Diff line number	Diff line
		@@ -215,4 +215,121 @@ public class DBinterface {
		}
		return new String();
		}

		public void updateTokenInDatabase(String sentence, DirectedGraph<State> graph){
		StateMachine SM = new StateMachine();
		sentence = sentence.replaceAll("\\p{Punct}", " $0");
		String[] tokens = sentence.split("\\s+");
		String[] tokensCopy = tokens.clone();
		List<Boolean> missFlag = new ArrayList();
		List<String> tokenList = new ArrayList<>(Arrays.asList(tokensCopy));
		String url = "jdbc:sqlite:./SQLite/mydatabase.db";
		String urlinsert = "jdbc:sqlite:./SQLite/newdatabase.db";
		String dicFileName = "./SQLite/smallDic.txt";
		TypoCorrector typoChecker = TypoCorrector.of(dicFileName);
		int initialConf = 0;
		int cntMiss = 0;
		try (Connection connection = DriverManager.getConnection(url)) {

		// Lookup each token in the database and categorize it
		for (int i = 0; i < tokens.length; i++) {
		String token = tokens[i];
		System.out.print("\nBefor token: "+tokens[i]+"\| ");
		try (Statement statement = connection.createStatement()) {

		String query = "SELECT role FROM word_roles WHERE word = '" + token + "';";
		String role = new String();

		ResultSet resultSet = statement.executeQuery(query);
		if (resultSet.next()) {
		role = resultSet.getString("role");
		//////System.out.print("first try: " + token + " -> " + role);
		System.out.println("role: " + role);
		if(!role.equals("NAN")){
		tokens[i] = role;
		missFlag.add(false);
		}else{
		missFlag.add(true);
		cntMiss ++;
		}
		}else{
		String tokenCorrected = new String();
		if(role.isEmpty()){
		tokenCorrected = typoChecker.closestWord(token);
		if(!tokenCorrected.equals(token))
		initialConf += 5;
		// ////System.out.print("Corrected token: " + token + " -> " + tokenCorrected);

		query = "SELECT role FROM word_roles WHERE word = '" + tokenCorrected + "';";
		// Replace the token with its role
		resultSet = statement.executeQuery(query);
		if (resultSet.next()) {
		tokenList.set(i,tokenCorrected);
		role = resultSet.getString("role");
		// ////System.out.print("\| Second try: "+ token + " -> " + role);
		System.out.println("role: " + role);
		if(!role.equals("NAN")){
		tokens[i] = role;
		missFlag.add(false);
		}else{
		missFlag.add(true);
		cntMiss ++;
		}
		}else{
		missFlag.add(true);
		cntMiss ++;
		}
		}

		}
		}
		System.out.print("After token: "+tokens[i]+"\| ");
		//////System.out.println();
		}
		System.out.println("\nMISS: "+cntMiss);
		if(cntMiss<3 && cntMiss>0){
		List<State> actions = new ArrayList<>();

		for(String token: tokens){
		actions.add(State.fromString(token));
		}
		// Define the initial state
		State initialState = State.START;

		// Check if the sequence of actions follows the state machine
		List<State> suggested = SM.updateDB(graph, actions, initialState);
		System.out.println(actions);
		System.out.println(suggested);
		System.out.println("---------------------------");
		int cntUp=0;
		for(int i=0; i<suggested.size(); i++){
		if(!suggested.get(i).toString().equals(tokens[i]))
		cntUp++;
		}
		System.out.println(missFlag);
		if(cntUp<3){
		for(int i=0; i<tokens.length; i++){
		if(missFlag.get(i)){
		if(!suggested.get(i).toString().equals(tokens[i])){
		System.out.println(tokens[i] + "\| " + suggested.get(i));
		String sql = "INSERT INTO word_roles VALUES(?, ?);";
		try (Connection conn = DriverManager.getConnection(urlinsert);
		PreparedStatement pstmt = conn.prepareStatement(sql)) {
		// Insert the first record
		pstmt.setString(1, tokens[i]);
		pstmt.setString(2, suggested.get(i).toString());
		pstmt.executeUpdate();

		} catch (SQLException e) {
		System.out.println(e.getMessage());
		}
		}
		}
		}
		}
		}
		} catch (SQLException e) {
		e.printStackTrace();
		}
		}
		}

CheckerCorrector/DirectedGraph/BasicGraph.java

+2 −0

Original line number	Diff line number	Diff line
		@@ -53,6 +53,7 @@ public class BasicGraph {
		graph.addEdge(State.NOT, State.ADVERB);
		graph.addEdge(State.NOT, State.ADJECTIVE);
		graph.addEdge(State.NOT, State.ARTICLE);
		graph.addEdge(State.NOT, State.DOT);
		graph.addEdge(State.OF, State.NOUN);
		graph.addEdge(State.NOUN, State.OF);
		graph.addEdge(State.NOUN, State.IS);
		@@ -60,6 +61,7 @@ public class BasicGraph {
		graph.addEdge(State.IS, State.ADJECTIVE);
		graph.addEdge(State.IS, State.ADVERB);
		graph.addEdge(State.IS, State.ARTICLE);
		graph.addEdge(State.IS, State.NOT);
		graph.addEdge(State.THAT, State.IF);

CheckerCorrector/DirectedGraph/DFS.java

+5 −5

Original line number	Diff line number	Diff line
		@@ -19,13 +19,13 @@ public class DFS {
		return new DFS();
		}

		public Set<String> dfs(DirectedGraph<State> graph, State currentState, int maxDepth){
		public Set<String> dfs(DirectedGraph<State> graph, State currentState, int maxDepth, int minDepth){
		List<State> path = new ArrayList<>();
		path.add(currentState);
		dfsRecurssion(graph, currentState, 0, maxDepth, path);
		dfsRecurssion(graph, currentState, 0, maxDepth, path, minDepth);
		return allPaths;
		}
		private void dfsRecurssion(DirectedGraph<State> graph, State currentState, int depth, int maxDepth, List<State> path) {
		private void dfsRecurssion(DirectedGraph<State> graph, State currentState, int depth, int maxDepth, List<State> path, int minDepth) {
		if ((currentState == State.DOT)) {
		ListToString lTS = ListToString.of();
		StringBuilder sb = new StringBuilder();
		@@ -34,7 +34,7 @@ public class DFS {
		lTS.addString(p);
		cnt++;
		}
		if(cnt>(maxDepth-4))
		if(cnt>(maxDepth-minDepth))
		allPaths.add(lTS.getString());
		return;
		}else if(depth >= maxDepth){
		@@ -43,7 +43,7 @@ public class DFS {
		List<State> transitions = graph.getAdjacentNodes(currentState);
		for (State nextState : transitions) {
		path.add(nextState);
		dfsRecurssion(graph, nextState, depth + 1, maxDepth, path);
		dfsRecurssion(graph, nextState, depth + 1, maxDepth, path, minDepth);
		path.remove(path.size() - 1);
		}
		}

CheckerCorrector/SQLite/mydatabase.db

(12 KiB)

File changed.

No diff preview for this file type.

View original file

View changed file