Further compress json data

.gitignore

@@ -5,3 +5,5 @@ es/
 lib/
 unicode-properties.js
 unicode-properties.min.js
+src/data.json
+src/trie.json

Makefile.js

@@ -19,12 +19,12 @@ target.generateTrieJson = () => {
   mkdir('lib')
   mkdir('dist')
   exec('babel-node src/generate.js');
-  cp('data.json', 'trie.json', 'src');
-  cp('data.json', 'trie.json', 'es');
-  cp('data.json', 'trie.json', 'lib');
+  cp('src/data.json', 'src/trie.json', 'es');
+  cp('src/data.json', 'src/trie.json', 'lib');
 };
 
 target.compileBabel = () => {
+  target.generateTrieJson();
   env.MODULE_TYPE = 'es6';
   exec(`babel src/index.js --out-dir es`);
   env.MODULE_TYPE = 'commonjs';
@@ -44,5 +44,5 @@ target.rollupUMDMin = () => {
 };
 
 target.clean = () => {
-  rm('-rf', 'data.json', 'trie.json', 'src/data.json', 'src/trie.json', 'dist', 'lib', 'es');
+  rm('-rf', 'src/data.json', 'src/trie.json', 'dist', 'lib', 'es');
 };

package.json

@@ -6,6 +6,8 @@
   "unpkg": "dist/unicode-properties.min.js",
   "module": "es/index.js",
   "dependencies": {
+    "base64-arraybuffer": "^0.1.5",
+    "pako": "^1.0.6",
     "unicode-trie": "^0.3.0"
   },
   "devDependencies": {

src/generate.js

@@ -1,6 +1,8 @@
 const codePoints = require('codepoints');
 const fs = require('fs');
 const UnicodeTrieBuilder = require('unicode-trie/builder');
+const pako = require('pako');
+const base64 = require('base64-arraybuffer');
 
 const log2 = Math.log2 || (n => Math.log(n) / Math.LN2);
 
@@ -96,11 +98,18 @@ for (codePoint of Array.from(codePoints)) {
 
 // Trie is serialized suboptimally as JSON so it can be loaded via require,
 // allowing unicode-properties to work in the browser
-fs.writeFileSync('./trie.json', JSON.stringify(trie.toBuffer()));
+const trieFilePath = __dirname + '/trie.json';
+const jsonBase64DeflatedTrie = JSON.stringify(base64.encode(pako.deflate(trie.toBuffer())));
+fs.writeFileSync(trieFilePath, jsonBase64DeflatedTrie);
 
-fs.writeFileSync('./data.json', JSON.stringify({
+const data = {
   categories: Object.keys(categories),
   combiningClasses: Object.keys(combiningClasses),
   scripts: Object.keys(scripts),
   eaw: Object.keys(eaws)
-}));
+};
+
+const dataFilePath = __dirname + '/data.json';
+const dataJsonBytes = JSON.stringify(data).split('').map(c => c.charCodeAt(0));
+const jsonBase64DeflatedData = JSON.stringify(base64.encode(pako.deflate(dataJsonBytes)));
+fs.writeFileSync(dataFilePath, jsonBase64DeflatedData);

src/index.js

@@ -1,8 +1,15 @@
 import UnicodeTrie from 'unicode-trie';
-import data from './data.json';
+import pako from 'pako';
+import * as base64 from 'base64-arraybuffer';
+
+import base64DeflatedData from './data.json';
+import base64DeflatedTrie from './trie.json';
+
+// Trie is serialized as a Buffer in node, but here
+// we may be running in a browser so we make an Uint8Array
+const data = JSON.parse(String.fromCharCode(...pako.inflate(base64.decode(base64DeflatedData))));
+const trieData = pako.inflate(base64.decode(base64DeflatedTrie));
 
-import trieBuffer from './trie.json';
-const trieData = new Uint8Array(trieBuffer.data);
 const trie = new UnicodeTrie(trieData);
 
 const log2 = Math.log2 || (n => Math.log(n) / Math.LN2);

yarn.lock

@@ -735,6 +735,11 @@ balanced-match@^1.0.0:
   version "1.0.0"
   resolved "https://registry.yarnpkg.com/balanced-match/-/balanced-match-1.0.0.tgz#89b4d199ab2bee49de164ea02b89ce462d71b767"
 
+base64-arraybuffer@^0.1.5:
+  version "0.1.5"
+  resolved "https://registry.yarnpkg.com/base64-arraybuffer/-/base64-arraybuffer-0.1.5.tgz#73926771923b5a19747ad666aa5cd4bf9c6e9ce8"
+  integrity sha1-c5JncZI7Whl0etZmqlzUv5xunOg=
+
 base@^0.11.1:
   version "0.11.2"
   resolved "https://registry.yarnpkg.com/base/-/base-0.11.2.tgz#7bde5ced145b6d551a90db87f83c558b4eb48a8f"
@@ -1975,6 +1980,11 @@ pako@^0.2.5:
   version "0.2.9"
   resolved "https://registry.yarnpkg.com/pako/-/pako-0.2.9.tgz#f3f7522f4ef782348da8161bad9ecfd51bf83a75"
 
+pako@^1.0.6:
+  version "1.0.7"
+  resolved "https://registry.yarnpkg.com/pako/-/pako-1.0.7.tgz#2473439021b57f1516c82f58be7275ad8ef1bb27"
+  integrity sha512-3HNK5tW4x8o5mO8RuHZp3Ydw9icZXx0RANAOMzlMzx7LVXhMJ4mo3MOBpzyd7r/+RUu8BmndP47LXT+vzjtWcQ==
+
 parse-glob@^3.0.4:
   version "3.0.4"
   resolved "https://registry.yarnpkg.com/parse-glob/-/parse-glob-3.0.4.tgz#b2c376cfb11f35513badd173ef0bb6e3a388391c"