addTypoCorrector (7ea368c0) · Commits · EC504 Spring 2024 Group Projects / Group6

DBinterface/DBinterface.java

+29 −4

Original line number	Diff line number	Diff line
		@@ -7,6 +7,7 @@ import DirectedGraph.DirectedGraph;

		import java.sql.*;
		import StateMachine.*;
		import TypoCorrector.TypoCorrector;

		public class DBinterface {
		public void checkTokenInDatabase(String sentence, DirectedGraph<State> graph){
		@@ -14,6 +15,9 @@ public class DBinterface {
		sentence = sentence.replaceAll("\\p{Punct}", " $0");
		String[] tokens = sentence.split("\\s+");
		String url = "jdbc:sqlite:./SQLite/mydatabase.db";
		String dicFileName = "./SQLite/smallDic.txt";
		TypoCorrector typoChecker = TypoCorrector.of(dicFileName);
		int initialConf = 0;
		try (Connection connection = DriverManager.getConnection(url)) {

		// Lookup each token in the database and categorize it
		@@ -21,14 +25,35 @@ public class DBinterface {
		String 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()) {
		String role = resultSet.getString("role");
		role = resultSet.getString("role");
		System.out.print("first try: " + token + " -> " + role);
		tokens[i] = role;
		}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()) {
		role = resultSet.getString("role");
		System.out.print("\| Second try: "+ token + " -> " + role);
		tokens[i] = role;
		}
		}

		}
		}
		System.out.println();
		}

		List<State> actions = new ArrayList<>();
		@@ -41,7 +66,7 @@ public class DBinterface {

		// Check if the sequence of actions follows the state machine

		int confidence = SM.isStateMachineFollowed(graph, actions, initialState);
		int confidence = SM.isStateMachineFollowed(graph, actions, initialState, initialConf);
		System.out.print("The confidence score is: "+ confidence + "\n");
		} catch (SQLException e) {
		e.printStackTrace();

SQLite/mydatabase.db

(12 KiB)

File changed.

No diff preview for this file type.

View original file

View changed file

SQLite/smallDic.txt

0 → 100644

+210687 −0

File added.

Preview size limit exceeded, changes collapsed.

StateMachine/StateMachine.java

+2 −2

Original line number	Diff line number	Diff line
		@@ -4,8 +4,8 @@ import java.util.List;

		import DirectedGraph.DirectedGraph;
		public class StateMachine{
		public int isStateMachineFollowed(DirectedGraph<State> graph, List<State> actions, State initialState) {
		int confidence = 0;
		public int isStateMachineFollowed(DirectedGraph<State> graph, List<State> actions, State initialState, int initialConf) {
		int confidence = initialConf;
		State currentState = initialState;
		for (State action : actions) {
		List<State> transitions = graph.getAdjacentNodes(currentState);

TypoCorrector/TypoCorrector.java

0 → 100644

+96 −0

Original line number	Diff line number	Diff line
		package TypoCorrector;

		import java.io.BufferedReader;
		import java.io.FileReader;
		import java.io.IOException;
		import java.util.ArrayList;

		public class TypoCorrector {
		TypoCorrector(String filePath){
		dic = new ArrayList<>();
		readWordsFromFile(filePath);
		}

		public static TypoCorrector of (String filename){
		return new TypoCorrector(filename);
		}
		public void readWordsFromFile(String filePath) {
		try (BufferedReader br = new BufferedReader(new FileReader(filePath))) {
		String line;
		while ((line = br.readLine()) != null) {
		line = line.replaceAll("[^a-zA-Z ]", "").toLowerCase(); // Remove non-alphabetic characters and convert to lowercase
		String[] words = line.split("\\s+"); // Split the line by spaces
		for (String word : words) {
		if (!word.isEmpty()) {
		dic.add(word); // Add the word to the dictionary
		}
		}
		}
		} catch (IOException e) {
		e.printStackTrace();
		}
		}

		public String closestWord(String word) {
		if(word.equals(".") \|\| word.equals(","))
		return word;
		int curMax = -(1<<30);
		String wordMax = new String();
		int indCur = 0;
		for(String dicS: dic){
		int m = dicS.length();
		int n = word.length();
		int[][] scoreMat = new int[m+1][n+1];
		int[][] dirMat = new int[m+1][n+1];

		for(int i=1; i<=n; i++){
		scoreMat[0][i] = (i-1)*gapExtend+gapOpen;
		}
		for(int i=1; i<=m; i++){
		scoreMat[i][0] = (i-1)*insertGapExtend+insertGapOpen;
		}
		for(int i=1; i<=m; i++){
		for(int j=1; j<=n; j++){
		int curScore = 0;
		int up = scoreMat[i][j-1] + ((dirMat[i][j-1]==2 \|\| dirMat[i][j-1]==0)?gapExtend:gapOpen);
		int left = scoreMat[i-1][j] + ((dirMat[i-1][j]==3 \|\| dirMat[i-1][j]==0)?insertGapExtend:insertGapOpen);
		int diag = scoreMat[i-1][j-1] + (word.charAt(j-1)==dicS.charAt(i-1)?matchScore:mismatchScore);


		curScore = diag;
		dirMat[i][j]= 1;

		if(up>curScore){
		curScore = up;
		dirMat[i][j]= 2;
		}
		if(left>curScore){
		dirMat[i][j]= 3;
		curScore = left;
		}
		scoreMat[i][j] = curScore;
		}
		}

		int disCur = scoreMat[m][n];
		if(disCur>curMax){
		wordMax = dicS;
		curMax = disCur;
		}
		indCur ++;
		}


		return (curMax<4 && wordMax.length()>0)?wordMax:word;

		}
		ArrayList<String> dic;
		final int mismatchScore = -2;
		final int matchScore = 0;
		final int gapOpen = -2;
		final int gapExtend = -1;
		final int insert = -1;
		final int insertGapOpen = -2;
		final int insertGapExtend = -1;
		final int minusInf = -(1<<4);
		}