Merge branch 'Reza' into 'master' (41f82c4e) · Commits · EC504 Spring 2024 Group Projects / Group6

CheckerCorrector/DBinterface/DBinterface.java

+36 −24

Original line number	Diff line number	Diff line
		@@ -15,6 +15,7 @@ import java.sql.*;
		import StateMachine.*;
		import TypoCorrector.FilePrefixComparator;
		import TypoCorrector.TypoCorrector;
		import TypoCorrector.WordCountSorter;
		import util.TwoListStruct;
		import util.StringFileWriter;
		import util.StringProcessor;
		@@ -236,15 +237,17 @@ public class DBinterface {
		if (resultSet.next()) {
		word = resultSet.getString("word");
		////System.out.println("Here I am: "+ word);
		if(i+biasToken-1>0){
		if(i+biasToken>0){
		String nextToken = word;
		String queryy = "SELECT role FROM word_roles WHERE word = '" + nextToken + "';";
		ResultSet resultSett = statement.executeQuery(queryy);
		if (resultSett.next()) {
		String roleNext = resultSett.getString("role");
		wordPairDatabase.bfsAndGetWords(tokenList.get(i+biasToken-1), 2);
		FilePrefixComparator findSimilarity = FilePrefixComparator.of("similarity_words.txt");
		String tokenCorrected = findSimilarity.findBestMatchingPrefix(nextToken);
		wordPairDatabase.bfsAndWriteCountsToFile(tokenList.get(i+biasToken-1), 2);
		WordCountSorter wordCountSorter = WordCountSorter.of("similarity_words.txt");
		List<String> possibleWords = wordCountSorter.getSortedWordsByCount();
		//System.out.println(possibleWords);
		for(String tokenCorrected: possibleWords){
		if(!nextToken.equals(tokenCorrected)){
		String queryNext = "SELECT role FROM word_roles WHERE word = '" + tokenCorrected + "';";
		ResultSet resultSetNext = statement.executeQuery(queryNext);
		@@ -252,6 +255,8 @@ public class DBinterface {
		String roleNextNext = resultSett.getString("role");
		if(roleNextNext.equals(roleNext)){
		word = tokenCorrected;
		break;
		}
		}
		}
		}
		@@ -272,7 +277,7 @@ public class DBinterface {

		}else{
		if(flagsCorrection.isEmpty()){
		sfw.appendString("(INSERTION INTO INDEX): "+ i + " -> "+ word + "*");
		sfw.appendString("(INSERTION INTO THE END) -> "+ word + "*");
		tokenList.add(word);
		}else if(flagsCorrection.get(flagsCorrectioncnt)){
		tokenList.add(word);
		@@ -306,15 +311,17 @@ public class DBinterface {
		if (resultSet.next()) {
		word = resultSet.getString("word");
		////System.out.println("Here I am: "+ word);
		if(i+biasToken-1>0){
		if(i+biasToken>0){
		String nextToken = word;
		String queryy = "SELECT role FROM word_roles WHERE word = '" + nextToken + "';";
		ResultSet resultSett = statement.executeQuery(queryy);
		if (resultSett.next()) {
		String roleNext = resultSett.getString("role");
		wordPairDatabase.bfsAndGetWords(tokenList.get(i+biasToken-1), 2);
		FilePrefixComparator findSimilarity = FilePrefixComparator.of("similarity_words.txt");
		String tokenCorrected = findSimilarity.findBestMatchingPrefix(nextToken);
		wordPairDatabase.bfsAndWriteCountsToFile(tokenList.get(i+biasToken-1), 2);
		WordCountSorter wordCountSorter = WordCountSorter.of("similarity_words.txt");
		List<String> possibleWords = wordCountSorter.getSortedWordsByCount();
		// System.out.println(possibleWords);
		for(String tokenCorrected: possibleWords){
		if(!nextToken.equals(tokenCorrected)){
		String queryNext = "SELECT role FROM word_roles WHERE word = '" + tokenCorrected + "';";
		ResultSet resultSetNext = statement.executeQuery(queryNext);
		@@ -322,16 +329,20 @@ public class DBinterface {
		String roleNextNext = resultSett.getString("role");
		if(roleNextNext.equals(roleNext)){
		word = tokenCorrected;
		break;
		}
		}
		}
		}
		}
		}
		if(flagsCorrection.isEmpty()){
		sfw.appendString("(INSERTION INTO THE END) -> "+ word + "*");

		if(i<tokenList.size()){
		sfw.appendString("(INSERTION INTO INDEX): "+ i + " -> "+ word + "*");
		tokenList.add(i+biasToken,word);
		}else{
		sfw.appendString("(INSERTION INTO THE END) -> "+ word + "*");
		tokenList.add(word);
		}
		}else if(flagsCorrection.get(flagsCorrectioncnt)){
		@@ -339,6 +350,7 @@ public class DBinterface {
		// System.out.println("here!");
		tokenList.add(i+biasToken,word);
		}else{
		//sfw.appendString("(INSERTION INTO THE END) -> "+ word + "*");
		tokenList.add(word);
		}
		}else{

CheckerCorrector/SQLite/token_database_english.db

(44 KiB)

File changed.

No diff preview for this file type.

View original file

View changed file

CheckerCorrector/SQLite/word_similarity_english.db

(52 KiB)

File changed.

No diff preview for this file type.

View original file

View changed file

CheckerCorrector/SimilarityCorrector/WordPairDatabase.java

+53 −0

Original line number	Diff line number	Diff line
		@@ -41,6 +41,7 @@ public class WordPairDatabase {
		while ((line = reader.readLine()) != null) {
		line = line.toLowerCase();
		line = StringProcessor.handleApostrophe(line);
		line = line.replaceAll("\\p{Punct}", " $0");
		String[] words = line.split("\\s+");
		for (int i = 0; i < words.length - 1; i++) {
		String word1 = words[i];
		@@ -133,6 +134,58 @@ public class WordPairDatabase {
		return nextWords;
		}

		public void bfsAndWriteCountsToFile(String startWord, int depth) {
		String outputFile = "similarity_words.txt";
		Map<String, Integer> wordCounts = new HashMap<>();
		Queue<String> queue = new ArrayDeque<>();
		queue.add(startWord);

		try (Connection conn = DriverManager.getConnection(url)) {
		while (!queue.isEmpty() && depth > 0) {
		int size = queue.size();
		for (int i = 0; i < size; i++) {
		String word = queue.poll();
		for (Map.Entry<String, Integer> entry : getCountsForWord(conn, word).entrySet()) {
		//String pair = word + " " + entry.getKey();
		wordCounts.put(entry.getKey(), entry.getValue());
		queue.add(entry.getKey()); // Add next word to queue for BFS
		}
		}
		depth--;
		}
		} catch (SQLException e) {
		System.err.println("Error accessing database: " + e.getMessage());
		}

		// Write word counts to the output file
		//System.out.print("WORDCOUNTS: ");
		//System.out.println(wordCounts);
		try (BufferedWriter writer = new BufferedWriter(new FileWriter(outputFile))) {
		for (Map.Entry<String, Integer> entry : wordCounts.entrySet()) {
		writer.write(entry.getKey() + " " + entry.getValue() + "\n");
		}
		} catch (IOException e) {
		System.err.println("Error writing to file: " + e.getMessage());
		}
		}

		private Map<String, Integer> getCountsForWord(Connection conn, String word) {
		Map<String, Integer> counts = new HashMap<>();
		String sql = "SELECT word2, count FROM word_pairs WHERE word1 = ?";

		try (PreparedStatement stmt = conn.prepareStatement(sql)) {
		stmt.setString(1, word);
		ResultSet rs = stmt.executeQuery();
		while (rs.next()) {
		counts.put(rs.getString("word2"), rs.getInt("count"));
		}
		} catch (SQLException e) {
		System.err.println("Error retrieving counts for word: " + e.getMessage());
		}

		return counts;
		}

		public static void main(String[] args) {
		String url = "word_pairs.db";
		String fileName = "sentences.txt";

CheckerCorrector/TypoCorrector/WordCountSorter.java

0 → 100644

+47 −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.*;

		public class WordCountSorter {
		private String filename;

		private WordCountSorter(String filename) {
		this.filename = filename;
		}

		public static WordCountSorter of(String filename){
		return new WordCountSorter(filename);
		}

		public List<String> getSortedWordsByCount() {
		Map<String, Integer> wordCountMap = new HashMap<>();
		try (BufferedReader reader = new BufferedReader(new FileReader(filename))) {
		String line;
		while ((line = reader.readLine()) != null) {
		String[] parts = line.split("\\s+");
		if (parts.length == 2) {
		String word = parts[0];
		int count = Integer.parseInt(parts[1]);
		wordCountMap.put(word, count);
		}
		}
		} catch (IOException e) {
		System.err.println("Error reading file: " + e.getMessage());
		}

		// Sort the map by value (count) in descending order
		List<Map.Entry<String, Integer>> sortedEntries = new ArrayList<>(wordCountMap.entrySet());
		sortedEntries.sort((e1, e2) -> e2.getValue().compareTo(e1.getValue()));

		// Extract sorted words from sorted entries
		List<String> sortedWords = new ArrayList<>();
		for (Map.Entry<String, Integer> entry : sortedEntries) {
		sortedWords.add(entry.getKey());
		}

		return sortedWords;
		}
		}