more generated code

Author: syjer

src/test/java/ch/digitalfondue/jfiveparse/DoubleArrayTrie.java

@@ -8,8 +8,8 @@
 
 // from https://medium.com/@lchang1994/deep-dive-dat-double-array-trie-f51e5e5f006c
 public class DoubleArrayTrie {
-    private int[] check;
-    private int[] base;
+    int[] check;
+    int[] base;
 
     public DoubleArrayTrie() {
         check = new int[128];
@@ -37,10 +37,6 @@ private void expand(int index) {
         base = Arrays.copyOf(base, newSize);
     }
 
-    public int charId(char c) {
-        return c >= 38 && c <= 122 ? entitiesCharIdx[c - 38] : -1;
-    }
-
     public void insert(String word) {
         int node = 1;
         for (int i = 0; i < word.length(); i++) {
@@ -163,8 +159,8 @@ private List<Integer> getChildrenOfOffset(int nodeIdx, int baseOffset) {
         if (baseOffset == 0) {
             return children;
         }
-        // 'a'-'z' maps to 1..26
-        for (int c = 1; c <= 26; c++) {
+        // the chars map to 1 up to MAX_CHAR_ID
+        for (int c = 1; c <= MAX_CHAR_ID; c++) {
             int idx = baseOffset + c;
             if (idx < check.length && check[idx] == nodeIdx) {
                 children.add(c);
@@ -212,8 +208,12 @@ void check() {
     }
 
 
+    // generated code
 
-
+    private static final int MAX_CHAR_ID = 62;
+    public int charId(char c) {
+        return c >= 38 && c <= 122 ? entitiesCharIdx[c - 38] : -1;
+    }
     private static final int[] entitiesCharIdx = new int[85];
     static {
         Arrays.fill(entitiesCharIdx, -1);

src/test/java/ch/digitalfondue/jfiveparse/GenerateEntities.java

@@ -17,6 +17,7 @@
 
 import com.google.gson.GsonBuilder;
 import com.google.gson.reflect.TypeToken;
+import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
 
 import java.io.ByteArrayOutputStream;
@@ -41,12 +42,11 @@ private static class EntityValues {
 
     // this generates the supporting array for mapping a character to a unique id
     // for the double array trie
+    // step 1
     @Test
+    @Disabled
     void generateCharactersMappingArray() throws IOException {
-        Type type = (new TypeToken<Map<String, EntityValues>>() {
-        }).getType();
-        String json = Files.readString(Paths.get("src/test/resources/entities.json"));
-        Map<String, EntityValues> m = new GsonBuilder().create().fromJson(json, type);
+        Map<String, EntityValues> m = getEntitiesMap();
 
         var allUsedChars = new TreeSet<Integer>();
         for (String key: m.keySet()) {
@@ -56,6 +56,7 @@ void generateCharactersMappingArray() throws IOException {
         var firstChar = allChars.get(0);
         var lastChar = allChars.get(allChars.size() - 1);
         // mapping function
+        System.out.println("private static final int MAX_CHAR_ID = " + allChars.size() + ";");
         System.out.println("public int charId(char c) {");
         System.out.printf("  return c >= %d && c <= %d ? entitiesCharIdx[c - %d] : -1;\n", firstChar, lastChar, firstChar);
         System.out.println("}");
@@ -70,14 +71,31 @@ void generateCharactersMappingArray() throws IOException {
         }
         System.out.println("}");
         //
-        System.out.println(allUsedChars.first() + " " + allUsedChars.last() + ";");
     }
 
-    public static void main(String[] args) throws IOException {
+    private static Map<String, EntityValues> getEntitiesMap() throws IOException {
         Type type = (new TypeToken<Map<String, EntityValues>>() {
         }).getType();
         String json = Files.readString(Paths.get("src/test/resources/entities.json"));
         Map<String, EntityValues> m = new GsonBuilder().create().fromJson(json, type);
+        return m;
+    }
+
+    // step 2
+    @Test
+    @Disabled
+    void pregenerateSupportArray() throws IOException {
+        var entities = getEntitiesMap();
+        var dat = new DoubleArrayTrie();
+        for(String key : entities.keySet()) {
+            dat.insert(key);
+        }
+        System.out.println(Arrays.toString(dat.base));
+        System.out.println(Arrays.toString(dat.check));
+    }
+
+    public static void main(String[] args) throws IOException {
+        Map<String, EntityValues> m = getEntitiesMap();
 
 
         // we can iterate for making a and array for mapping char -> id!