Code after lecture

01/Main.java

+import java.io.FileNotFoundException;
+
+public class Main {
+    public static char getBestChoice(QuadGramLikelihoods likelihoods, String before, String after) {
+        char bestLetter = '?';
+        double best = Double.NEGATIVE_INFINITY;
+        /*
+        char bestLetter = 'A';
+        double best = likelihoods.get(before + 'A' + after);
+        */
+        for (char letter = 'A'; letter <= 'Z'; letter++) {
+            String quadgram = before + letter + after;
+            if (likelihoods.get(quadgram) > best) {
+                bestLetter = letter;
+                best = likelihoods.get(quadgram);
+            }
+        }
+        return bestLetter;
+        // for every letter in the alphabet,
+            // combine before + L + after
+            // ask for for likelihood of that combo
+            // keep a running "maximum" letter
+    }
+
+    private static final int N = 4;
+
+    public static String[] deduceHiddenLetter(QuadGramLikelihoods likelihoods, String hidden) {
+        String[] results = new String[N];
+        int qmark = hidden.indexOf('?');
+        for (int i = 0; i < N; i++) {
+            String before = hidden.substring(qmark - i, qmark);
+            String after = hidden.substring(qmark + 1, qmark - i + N);
+            System.out.println(before + "?" + after);
+            char best = getBestChoice(likelihoods, before, after);
+            results[i] = hidden.substring(0, qmark) +
+                    best +
+                    hidden.substring(qmark + 1, hidden.length());
+        }
+        return results;
+    }
+
+    public static void main(String[] args) throws FileNotFoundException {
+        QuadGramLikelihoods likelihoods = new QuadGramLikelihoods();
+
+        String hidden = "PSYC?OLOGY";
+        //SYC? or YC?O ...
+        String[] deduced = deduceHiddenLetter(likelihoods, hidden);
+        for (int i = 0; i < deduced.length; i++) {
+            System.out.println(hidden + " -> " + deduced[i]);
+        }
+    }
+}

01/QuadGramLikelihoods.java

+import java.io.File;
+import java.io.FileNotFoundException;
+import java.util.HashMap;
+import java.util.Map;
+import java.util.Scanner;
+
+public class QuadGramLikelihoods {
+    private Map<String, Double> likelihoods;
+    private double notFoundLikelihood;
+
+    public QuadGramLikelihoods() throws FileNotFoundException {
+        Map<String, Integer> freqs = getQuadFrequencies();
+
+        long total = 0;
+        for (String qg : freqs.keySet()) {
+            total += freqs.get(qg);
+        }
+
+        total += freqs.keySet().size();
+
+        this.likelihoods = new HashMap<>();
+        for (String qg : freqs.keySet()) {
+            this.likelihoods.put(qg, Math.log10((double)freqs.get(qg) / (double)total));
+        }
+        this.notFoundLikelihood = Math.log10(1.0/total);
+    }
+
+    public static Map<String, Integer> getQuadFrequencies() throws FileNotFoundException {
+        Scanner quadgrams = new Scanner(new File("english_quadgrams.txt"));
+        Map<String, Integer> freqs = new HashMap<>();
+
+        while (quadgrams.hasNextLine()) {
+            String line = quadgrams.nextLine();
+            String[] pieces = line.split("\\s");
+            String quadgram = pieces[0];
+            int freq = Integer.parseInt(pieces[1]);
+            freqs.put(quadgram, freq);
+        }
+
+        return freqs;
+    }
+
+    /**
+     * Returns the log-likelihood of the provided quad-gram.
+     * @param quadgram the quad-gram to get the likelihood of
+     * @return the log-likelihood as calculated by our data set
+     */
+    public double get(String quadgram) {
+        Double likelihood = likelihoods.get(quadgram);
+        if (likelihood != null) {
+            return likelihood;
+        }
+        else {
+            return this.notFoundLikelihood;
+        }
+    }
+}

01/Tests.java

+import org.junit.jupiter.api.*;
+import org.junit.jupiter.params.ParameterizedTest;
+import org.junit.jupiter.params.provider.CsvSource;
+
+import java.io.FileNotFoundException;
+
+import static org.junit.jupiter.api.Assertions.*;
+
+
+public class Tests {
+    @Order(0)
+    @DisplayName("Test deduceHiddenLetter")
+    @ParameterizedTest
+    @CsvSource({
+            "?SYCHOLOGY,,,,PSYCHOLOGY",
+            "P?YCHOLOGY,,,PSYCHOLOGY,PSYCHOLOGY",
+            "PS?CHOLOGY,,PSYCHOLOGY, PSUCHOLOGY, PSSCHOLOGY",
+            "PSY?HOLOGY,PSYCHOLOGY, PSYCHOLOGY, PSYSHOLOGY, PSYWHOLOGY",
+            "PSYC?OLOGY,PSYCHOLOGY, PSYCHOLOGY, PSYCHOLOGY, PSYCCOLOGY",
+            "PSYCH?LOGY,PSYCHALOGY, PSYCHILOGY, PSYCHELOGY, PSYCHOLOGY",
+            "PSYCHO?OGY,PSYCHOOOGY, PSYCHOTOGY, PSYCHOLOGY, PSYCHOLOGY",
+            "PSYCHOL?GY,PSYCHOLDGY, PSYCHOLOGY, PSYCHOLOGY,",
+            "PSYCHOLO?Y,PSYCHOLOGY, PSYCHOLOGY,,",
+            "PSYCHOLOG?,PSYCHOLOGY,,,",
+            "PSYCHOLOGY,PSYCHOLOGY,PSYCHOLOGY,PSYCHOLOGY,PSYCHOLOGY",
+    })
+    public void testDeduceHiddenLetter(String hidden, String correct0, String correct1, String correct2, String correct3) throws FileNotFoundException {
+        QuadGramLikelihoods likelihoods = new QuadGramLikelihoods();
+        String[] correct = {correct0, correct1, correct2, correct3};
+
+        String[] deduced = Main.deduceHiddenLetter(likelihoods, hidden);
+        assertArrayEquals(correct, deduced);
+
+    }
+}