thurtle: Add lzw & ulz compression

src/web/thurtle/lzw.ts

@@ -0,0 +1,46 @@
+export function lzwEncode(data: Uint8Array) {
+  const dict = new Map<string, number>();
+  const out: Array<number> = [];
+  let phrase = String.fromCharCode(data[0]);
+  let code = 256;
+  for (let i = 1; i < data.length; i++) {
+    const chr = String.fromCharCode(data[i]);
+    const nphrase = phrase + chr;
+    if (dict.has(nphrase)) {
+      phrase = nphrase;
+    } else {
+      out.push(phrase.length > 1 ? dict.get(phrase)! : phrase.charCodeAt(0));
+      dict.set(nphrase, code);
+      code++;
+      phrase = chr;
+    }
+  }
+  out.push(phrase.length > 1 ? dict.get(phrase)! : phrase.charCodeAt(0));
+  return new TextEncoder().encode(
+    out.map((c) => String.fromCharCode(c)).join("")
+  );
+}
+
+export function lzwDecode(rdata: Uint8Array) {
+  const data = [...new TextDecoder().decode(rdata)].map((c) => c.charCodeAt(0));
+  let dict = new Map<number, number[]>();
+  var curChar = data[0];
+  var curPhrase = [curChar];
+  var out = [curChar];
+  var code = 256;
+  for (var i = 1; i < data.length; i++) {
+    const c = data[i];
+    const phrase =
+      c < 256
+        ? [data[i]]
+        : dict.has(c)
+        ? dict.get(c)!
+        : curPhrase.concat(curChar);
+    out.push(...phrase);
+    curChar = phrase[0];
+    dict.set(code, curPhrase.concat(curChar));
+    code++;
+    curPhrase = phrase;
+  }
+  return new Uint8Array(out);
+}

src/web/thurtle/ulz.ts

@@ -0,0 +1,104 @@
+export function ulzDecode(src: Uint8Array) {
+  const dst: Array<number> = [];
+  let sp = 0;
+  while (sp < src.length) {
+    const c = src[sp++];
+    if (c & 0x80) {
+      // CPY
+      let length;
+      if (c & 0x40) {
+        if (sp >= src.length) {
+          throw new Error(`incomplete CPY2`);
+        }
+        length = ((c & 0x3f) << 8) | src[sp++];
+      } else {
+        length = c & 0x3f;
+      }
+      if (sp >= src.length) {
+        throw new Error(`incomplete CPY`);
+      }
+      let cp = dst.length - (src[sp++] + 1);
+      if (cp < 0) {
+        throw new Error(`CPY underflow`);
+      }
+      for (let i = 0; i < length + 4; i++) {
+        dst.push(dst[cp++]);
+      }
+    } else {
+      // LIT
+      if (sp + c >= src.length) {
+        throw new Error(`LIT out of bounds: ${sp} + ${c} >= ${src.length}`);
+      }
+      for (let i = 0; i < c + 1; i++) {
+        dst.push(src[sp++]);
+      }
+    }
+  }
+  return new Uint8Array(dst);
+}
+
+const MIN_MAX_LENGTH = 4;
+
+function findBestMatch(
+  src: Uint8Array,
+  sp: number,
+  dlen: number,
+  slen: number
+) {
+  let bmlen = 0;
+  let bmp = 0;
+  let dp = sp - dlen;
+  for (; dlen; dp++, dlen--) {
+    let i = 0;
+    for (; ; i++) {
+      if (i == slen) {
+        return [dp, i];
+      }
+      if (src[sp + i] != src[dp + (i % dlen)]) {
+        break;
+      }
+    }
+    if (i > bmlen) {
+      bmlen = i;
+      bmp = dp;
+    }
+  }
+  return [bmp, bmlen];
+}
+
+export function ulzEncode(src: Uint8Array) {
+  let dst: Array<number> = [];
+  let sp = 0;
+  let litp = -1;
+  while (sp < src.length) {
+    const dlen = Math.min(sp, 256);
+    const slen = Math.min(src.length - sp, 0x3fff + MIN_MAX_LENGTH);
+    const [bmp, bmlen] = findBestMatch(src, sp, dlen, slen);
+    if (bmlen >= MIN_MAX_LENGTH) {
+      // CPY
+      const bmctl = bmlen - MIN_MAX_LENGTH;
+      if (bmctl > 0x3f) {
+        // 	CPY2
+        dst.push((bmctl >> 8) | 0xc0);
+        dst.push(bmctl & 0xff);
+      } else {
+        dst.push(bmctl | 0x80);
+      }
+      dst.push(sp - bmp - 1);
+      sp += bmlen;
+      litp = -1;
+    } else {
+      // LIT
+      if (litp >= 0) {
+        if ((dst[litp] += 1) == 127) {
+          litp = -1;
+        }
+      } else {
+        dst.push(0);
+        litp = dst.length - 1;
+      }
+      dst.push(src[sp++]);
+    }
+  }
+  return new Uint8Array(dst);
+}