Commit 274231a2 authored by Manuel Segimon's avatar Manuel Segimon
Browse files

Refactor Corrector and Checker classes to accept metadata file path as a parameter

parent dd38b617

src/main/java/edu/bu/LanguageCorrection/Checker.java

+6 −5
Original line number Diff line number Diff line
@@ -11,10 +11,10 @@ import java.io.ByteArrayOutputStream; 
import java.io.FileInputStream;



public class Checker {

    public void analyze(String text) {

    public void analyze(String text, String metadataPath) {

        List<String> sentences = TextProcessor.extractSentences(text);



        TrieNode detector = loadFile("metadata.ser");

        TrieNode detector = loadFile(metadataPath);



        Map<String, Float> sentenceScores = new HashMap<>();

        Map<String, Float> phraseScores = new HashMap<>();
@@ -97,14 +97,15 @@ public class Checker { 
    }



    public static void main(String[] args) {

        if (args.length > 1 && "--file".equals(args[0])) { // check syntax

            String path = args[1];

        if (args.length > 3 && "--LF".equals(args[0]) && "--file".equals(args[2])) { // check syntax

            String metadataPath = args[1];

            String path = args[3];

            try {

                // Read entire file

                String content = new String(Files.readAllBytes(Paths.get(path)));



                Checker checker = new Checker(); // Run checker

                checker.analyze(content);

                checker.analyze(content, metadataPath);

            } catch (IOException e) {

                System.err.println("Error reading file: " + e.getMessage());

            }

src/main/java/edu/bu/LanguageCorrection/Corrector.java

+6 −5
Original line number Diff line number Diff line
@@ -17,8 +17,8 @@ import java.util.HashMap; 
public class Corrector {

    private TrieNode detector;



    public Corrector() {

        detector = loadFile("metadata.ser");

    public Corrector(String metadataPath) {

        detector = loadFile(metadataPath);

    }



    public TrieNode getDetector() {
@@ -209,11 +209,12 @@ public class Corrector { 
    }



    public static void main(String[] args) {

        if (args.length > 1 && "--file".equals(args[0])) {

            String path = args[1];

        if (args.length > 3 && "--LF".equals(args[0]) && "--file".equals(args[2])) {

            String metadataPath = args[1];

            String path = args[3];

            try {

                String content = new String(Files.readAllBytes(Paths.get(path)));

                Corrector corrector = new Corrector(); // Run corrector

                Corrector corrector = new Corrector(metadataPath); // Run corrector

                String[] sentences = TextProcessor.extractSentences(content).toArray(new String[0]);

                for (String sentence : sentences) {

                    sentence = sentence.replaceAll("[^a-zA-Z0-9\\s]", "");

src/main/java/edu/bu/LanguageCorrection/MainApp.java

+30 −5
Original line number Diff line number Diff line
@@ -11,6 +11,7 @@ import java.awt.*; 
// import java.awt.event.ActionEvent;

import java.io.ByteArrayOutputStream;

import java.io.PrintStream;

import java.io.StringBufferInputStream;

import java.nio.file.Files;

import java.nio.file.Paths;

import java.util.List;
@@ -33,8 +34,11 @@ public class MainApp extends JFrame { 
    private JTextArea resultArea;

    private Highlighter.HighlightPainter myHighlightPainter;

    private final JButton runButton;

    private final JButton changeLanguageButton;

    private final JComboBox<String> moduleSelector;



    private String languageFile = "English.ser";



    public MainApp() {

        super("Language Correction Tool");

        setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);
@@ -43,11 +47,12 @@ public class MainApp extends JFrame { 
        setLayout(new BorderLayout());



        // Check if metadata file exists

        if (!Files.exists(Paths.get("metadata.ser"))) {

        if (!Files.exists(Paths.get(languageFile))) {

            String[] languages = {"English", "German", "Portuguese", "Italian"};

            String selectedLanguage = (String) JOptionPane.showInputDialog(this, "Metadata file not found. Please choose a language to build off of. \n(If you want to build from scratch just click Cancel)", "Language Selection", JOptionPane.PLAIN_MESSAGE, null, languages, languages[0]);

            if (selectedLanguage != null) {

                crawler webCrawler = new crawler();

                languageFile = selectedLanguage + ".ser";

                crawler webCrawler = new crawler(languageFile);

                webCrawler.build(selectedLanguage);

            }

        }
@@ -86,10 +91,30 @@ public class MainApp extends JFrame { 
                    JOptionPane.showMessageDialog(this, "Select a valid module");

            }

        });

        changeLanguageButton = new JButton("Change Language");

        changeLanguageButton.addActionListener(e -> {

            String[] languages = {"English", "German", "Portuguese", "Italian"};

            String selectedLanguage = (String) JOptionPane.showInputDialog(this, "Please choose a language to build off of.", "Language Selection", JOptionPane.PLAIN_MESSAGE, null, languages, languages[0]);

            if (selectedLanguage != null) {

                languageFile = selectedLanguage + ".ser";



                // Check if metadata file exists

                String extraText = "";

                if (!Files.exists(Paths.get(languageFile))) {

                    crawler webCrawler = new crawler(languageFile);

                    webCrawler.build(selectedLanguage);

                    extraText = "New metadata file created.\n";

                }



                JOptionPane.showMessageDialog(this, extraText + "Language changed to " + selectedLanguage + ".");

            }

        });



        JPanel northPanel = new JPanel(new BorderLayout());

        northPanel.add(moduleSelector, BorderLayout.WEST);

        northPanel.add(urlField, BorderLayout.CENTER);

        northPanel.add(runButton, BorderLayout.EAST);

        northPanel.add(changeLanguageButton, BorderLayout.NORTH);



        // Result area

        resultArea = new JTextArea();
@@ -102,7 +127,7 @@ public class MainApp extends JFrame { 


    private void runCrawler(String input) {

        StringBuilder outputBuilder = new StringBuilder();

        crawler webCrawler = new crawler(output -> {

        crawler webCrawler = new crawler(languageFile, output -> {

            outputBuilder.append(output);

            SwingUtilities.invokeLater(() -> resultArea.setText(outputBuilder.toString()));

        });
@@ -150,7 +175,7 @@ public class MainApp extends JFrame { 


            for (String sentence : sentences) {

                System.setOut(printStream);

                checker.analyze(sentence); // Analyze each sentence separately

                checker.analyze(sentence, languageFile); // Analyze each sentence separately



                // Reset System.out

                System.out.flush();
@@ -270,7 +295,7 @@ public class MainApp extends JFrame { 
                content= new String(Files.readAllBytes(Paths.get(input)));

            else

                content = input;

            Corrector corrector = new Corrector();

            Corrector corrector = new Corrector(languageFile);

            String[] sentences = TextProcessor.extractSentences(content).toArray(new String[0]);



            StringBuilder result = new StringBuilder();

src/main/java/edu/bu/LanguageCorrection/crawler.java

+17 −5
Original line number Diff line number Diff line
@@ -26,12 +26,22 @@ import java.lang.Math; 
public class crawler {



    public static void main(String[] args) throws IOException {

        // Extract metadata file from first argument

        String file;

        if (!args[0].equals("--LF")) {

            System.err.println("Invalid argument. Please provide the metadata file.");

            return;

        } else {

            file = args[1];

        }





        // Initialize web crawler

        crawler web_crawler = new crawler();

        crawler web_crawler = new crawler(file);



        // Open the file and read all lines (URLs) into the queue

        String file_url = "";

        for (int i = 0; i < args.length; i++) {

        for (int i = 2; i < args.length; i++) {

            if (args[i].equals("--file")) {

                file_url = args[i + 1];

            } else if (args[i].equals("--debug")) {
@@ -80,7 +90,7 @@ public class crawler { 
    // members

    private final LinkedList<String> url_queue;

    private final HashSet<String> visited_urls;

    private String filePath = "metadata.ser";

    private String filePath;

    private TrieNode wordUsage = new TrieNode();

    private int compression_size = 0;

    private boolean debug = false; // flag that outputs uncompressed json showing the trie
@@ -89,23 +99,25 @@ public class crawler { 
    private static final int MAXNGRAM = 3;

    private Consumer<String> outputCallback;



    public crawler() {

    public crawler(String file) {

        url_queue = new LinkedList<>();

        visited_urls = new HashSet<>();



        // Load trie from file

        filePath = file;

        wordUsage = loadFile(filePath);

        this.outputCallback = System.out::println;



        // Estimate page count based on compressed file size

    }



    public crawler(Consumer<String> outputCallback) {

    public crawler(String file, Consumer<String> outputCallback) {

        this.outputCallback = outputCallback;

        url_queue = new LinkedList<>();

        visited_urls = new HashSet<>();



        // Load trie from file

        filePath = file;

        wordUsage = loadFile(filePath);

    }