Optimized the compacting algorithm

It now uses hashes to find duplicates Added status message for the compacting algorithm
NoaSenesi · Oct 31, 2023 · 330c47b · 330c47b
1 parent 9b4c797
commit 330c47b
Show file tree

Hide file tree

Showing 6 changed files with 57 additions and 91 deletions.
diff --git a/mkpkg b/mkpkg
@@ -16,7 +16,7 @@ echo "#!/bin/sh" >> package/usr/local/bin/g2t
 echo java -jar /usr/local/bin/g2t.jar \$@ >> package/usr/local/bin/g2t
 
 echo Package: g2t >> package/DEBIAN/control
-echo Version: 3.0.3 >> package/DEBIAN/control
+echo Version: 3.1.0 >> package/DEBIAN/control
 echo Maintainer: Noa Senesi >> package/DEBIAN/control
 echo Architecture: all >> package/DEBIAN/control
 echo Description: Grammar2Table >> package/DEBIAN/control

diff --git a/src/fr/senesi/g2t/Grammar2Table.java b/src/fr/senesi/g2t/Grammar2Table.java
@@ -12,7 +12,7 @@
 import fr.senesi.g2t.tokenizer.Tokenizer;
 
 public class Grammar2Table {
-	public static final String VERSION = "3.0.3";
+	public static final String VERSION = "3.1.0";
 
 	public static void main(String[] args) {
 		if (args.length == 0) {

diff --git a/src/fr/senesi/g2t/fsm/FiniteStateMachine.java b/src/fr/senesi/g2t/fsm/FiniteStateMachine.java
@@ -1,6 +1,7 @@
 package fr.senesi.g2t.fsm;
 
 import java.util.ArrayList;
+import java.util.Collections;
 import java.util.LinkedHashSet;
 import java.util.List;
 
@@ -115,35 +116,40 @@ public void createAllStates() {
 
 	public List<List<State>> findDuplicates() {
 		List<List<State>> duplicates = new ArrayList<>();
-		List<State> visited = new ArrayList<>();
+
+        List<State> sortedStates = new ArrayList<>(states);
 
-		for (int i = 0; i < states.size() - 1; i++) {
-			State s1 = getStates().get(i);
-			if (visited.contains(s1)) continue;
+        Collections.sort(sortedStates, (s1, s2) -> {
+            if (s1.coreEquals(s2)) return 0;
 
-			List<State> duplicate = new ArrayList<>();
+            return s1.coreHashCode() < s2.coreHashCode() ? -1 : 1;
+        });
 
-			duplicate.add(s1);
-			visited.add(s1);
+        for (int i = 1, start = 0; i < states.size(); i++) {
+            while (sortedStates.get(start).coreEquals(sortedStates.get(i))) {
+                i++;
 
-			for (int j = i + 1; j < states.size(); j++) {
-				State s2 = getStates().get(j);
+                if (i >= states.size()) {
+                    //i = states.size() - 1;
+                    break;
+                }
+            }
 
-				if (visited.contains(s2)) continue;
+            if (i - start <= 1) {
+                start = i;
+                continue;
+            }
 
-				if (s1.coreEquals(s2)) {
-					duplicate.add(s2);
-					visited.add(s2);
-				}
-			}
+            List<State> group = new ArrayList<>();
+            while (start < i) group.add(sortedStates.get(start++));
 
-			if (duplicate.size() > 1) duplicates.add(duplicate);
-		}
+            duplicates.add(group);
+        }
 
 		return duplicates;
 	}
 
 	public void setQuiet(boolean quiet) {
 		this.quiet = quiet;
 	}
-}
+}
diff --git a/src/fr/senesi/g2t/fsm/Rule.java b/src/fr/senesi/g2t/fsm/Rule.java
@@ -104,6 +104,13 @@ public int hashCode() {
 		return code * right.size() * context.size();
 	}
 
+	public int coreHashCode() {
+		int code = left.hashCode() * cursor;
+		for (String s : right) code += s.hashCode();
+
+		return code * right.size();
+	}
+
 	@Override
 	public boolean equals(Object o) {
 		if (!(o instanceof Rule)) return false;
@@ -127,4 +134,4 @@ public Rule copy() {
 
 		return rule;
 	}
-}
+}
diff --git a/src/fr/senesi/g2t/fsm/State.java b/src/fr/senesi/g2t/fsm/State.java
@@ -98,6 +98,14 @@ public int hashCode() {
 		return code * rules.size();
 	}
 
+    public int coreHashCode() {
+        int code = 0;
+
+        for (Rule r : rules) code += r.coreHashCode();
+
+        return code * rules.size();
+    }
+
 	@Override
 	public boolean equals(Object o) {
 		if (!(o instanceof State)) return false;
@@ -164,4 +172,4 @@ public void print() {
 	public String toString() {
 		return "State: I" + id;
 	}
-}
+}
diff --git a/src/fr/senesi/g2t/table/Table.java b/src/fr/senesi/g2t/table/Table.java
@@ -97,6 +97,8 @@ private void separateConflicts(List<List<State>> duplicates) {
 		Action[][] table = getTable();
 		List<List<State>> last = new ArrayList<>();
 
+        float compactingRatio = 1;
+
 		while (!duplicates.equals(last)) {
 			last = new ArrayList<>(duplicates);
 
@@ -139,8 +141,18 @@ private void separateConflicts(List<List<State>> duplicates) {
 
 				if (group.size() == 1) duplicates.remove(i--);
 			}
+
+            if (quiet) continue;
+
+            int duplicateCount = -duplicates.size();
+            for (List<State> duplicate : duplicates) duplicateCount += duplicate.size();
+            compactingRatio = (float) fsm.getStates().size() / (float) (fsm.getStates().size() - duplicateCount);
+
+            System.out.print("Current compacting ratio: " + String.format("%.2f", compactingRatio) + "\r");
 		}
-	}
+
+        if (!quiet) System.out.println("Final compacting ratio: " + String.format("%.2f", compactingRatio) + "          ");
+    }
 
 	private void mergeLines(Action[] line1, Action[] line2) {
 		// Lines are supposedly non-conflictual
@@ -219,73 +231,6 @@ public void compact() {
 		updateStateCount(duplicates);
 	}
 
-	public void print() {
-		int[] width = new int[tokens.size() + 1];
-		width[0] = 5;
-
-		for (int i = 1; i < width.length; i++) {
-			if (tokens.get(i-1).length() > width[i]) width[i] = tokens.get(i-1).length();
-		}
-
-		for (Action[] line : getTable()) {
-			for (int i = 0; i < line.length; i++) {
-				if (line[i].toString().length() > width[i+1]) width[i+1] = line[i].toString().length();
-			}
-
-			for (State s : fsm.getStates()) {
-				if (("I" + s.getId()).length() > width[0]) width[0] = ("I" + s.getId()).length();
-			}
-		}
-
-		System.out.print("┌");
-		for (int i = 0; i < width.length; i++) {
-			for (int j = 0; j < width[i] + 2; j++) System.out.print("─");
-			if (i != width.length - 1) System.out.print("┬");
-			if (i == fsm.getGrammar().getTerminals().size() + 1 || i == 0) System.out.print("┬");
-		}
-		System.out.println("┐");
-
-		System.out.print("│ State ");
-		for (int i = 0; i < width[0] - 5; i++) System.out.print(" ");
-		System.out.print("│");
-		for (int i = 1; i < width.length; i++) {
-			System.out.print("│ ");
-			System.out.print(tokens.get(i-1));
-			for (int j = 0; j < width[i] - tokens.get(i-1).length() + 1; j++) System.out.print(" ");
-			if (i == fsm.getGrammar().getTerminals().size() + 1) System.out.print("│");
-		}
-		System.out.println("│");
-
-		System.out.print("├");
-		for (int i = 0; i < width.length; i++) {
-			for (int j = 0; j < width[i] + 2; j++) System.out.print("─");
-			if (i != width.length - 1) System.out.print("┼");
-			if (i == fsm.getGrammar().getTerminals().size() + 1 || i == 0) System.out.print("┼");
-		}
-		System.out.println("┤");
-
-		for (int i = 0; i < getTable().length; i++) {
-			System.out.print("│ I" + i + " ");
-			for (int j = 0; j < width[0] - ("I" + i).length(); j++) System.out.print(" ");
-			System.out.print("│");
-			for (int j = 0; j < getTable()[i].length; j++) {
-				System.out.print("│ ");
-				System.out.print(getTable()[i][j]);
-				for (int k = 0; k < width[j+1] - getTable()[i][j].toString().length() + 1; k++) System.out.print(" ");
-				if (j == fsm.getGrammar().getTerminals().size()) System.out.print("│");
-			}
-			System.out.println("│");
-		}
-
-		System.out.print("└");
-		for (int i = 0; i < width.length; i++) {
-			for (int j = 0; j < width[i] + 2; j++) System.out.print("─");
-			if (i != width.length - 1) System.out.print("┴");
-			if (i == fsm.getGrammar().getTerminals().size() + 1 || i == 0) System.out.print("┴");
-		}
-		System.out.println("┘");
-	}
-
 	public void save(String filename) {
 		File file = new File(filename);
 		if (file.exists()) file.delete();
@@ -431,4 +376,4 @@ else if (action.getType() == ActionType.REDUCE) {
 			e.printStackTrace();
 		}
 	}
-}
+}