From 5804b02efca5646c0161ab5bff8a95fa820d269b Mon Sep 17 00:00:00 2001
From: ted chang <htchang@us.ibm.com>
Date: Mon, 30 Dec 2019 06:59:37 -0800
Subject: [PATCH 1/7] Initial commit for Text-sentiment-classifier

---
 bert-text-classifier/README.md                | 121 ++++++++++++
 bert-text-classifier/examples/test.html       |  14 ++
 bert-text-classifier/examples/test.js         |   2 +
 bert-text-classifier/package.json             |  82 ++++++++
 bert-text-classifier/rollup.config.js         |  78 ++++++++
 .../src/js/sentimentanalysis.ts               |  51 +++++
 .../src/server/sentimentanalysis.ts           |  62 ++++++
 bert-text-classifier/src/server/util.ts       |  29 +++
 bert-text-classifier/src/test.ts              |  79 ++++++++
 .../src/text-sentiment-classifier.ts          |  62 ++++++
 bert-text-classifier/src/tokenization.ts      | 184 ++++++++++++++++++
 bert-text-classifier/tsconfig.json            |  27 +++
 12 files changed, 791 insertions(+)
 create mode 100644 bert-text-classifier/README.md
 create mode 100644 bert-text-classifier/examples/test.html
 create mode 100644 bert-text-classifier/examples/test.js
 create mode 100644 bert-text-classifier/package.json
 create mode 100644 bert-text-classifier/rollup.config.js
 create mode 100644 bert-text-classifier/src/js/sentimentanalysis.ts
 create mode 100644 bert-text-classifier/src/server/sentimentanalysis.ts
 create mode 100644 bert-text-classifier/src/server/util.ts
 create mode 100644 bert-text-classifier/src/test.ts
 create mode 100644 bert-text-classifier/src/text-sentiment-classifier.ts
 create mode 100644 bert-text-classifier/src/tokenization.ts
 create mode 100644 bert-text-classifier/tsconfig.json

diff --git a/bert-text-classifier/README.md b/bert-text-classifier/README.md
new file mode 100644
index 0000000..637cc64
--- /dev/null
+++ b/bert-text-classifier/README.md
@@ -0,0 +1,121 @@
+# MAX for TensorFlow.js: Text Sentiment Classifier
+
+This is a TensorFlow.js port of the [MAX Human Pose Estimator](https://developer.ibm.com/exchanges/models/all/max-text-sentiment-classifier/) This model is able to detect whether a text fragment leans towards a positive or a negative sentiment.
+
+## Install
+
+### Browser
+
+```html
+<script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs"></script>
+<script src="https://cdn.jsdelivr.net/npm/@codait/text-sentiment-classifier"></script>
+```
+
+### Node.js
+
+```
+npm install --save @codait/text-sentiment-classifier
+```
+
+## Usage
+
+The complete examples for browser and Node.js environments are in the [`/examples`](https://github.com/CODAIT/max-tfjs-models/tree/master/text-sentiment-classifier/examples) directory.
+
+### Browser
+
+> **Note**: _When loaded in a browser, the global variable `textSentimentClassifier` will be available to access the API._
+
+```javascript
+
+textSentimentClassifier
+  .predict("i like strawberries")
+  .then(prediction => {
+    console.log(prediction)
+  });
+```
+
+### Node.js
+
+```javascript
+const tc = require('@codait/text-sentiment-classifier');
+tc.predict("i like strawberries").then(res=>console.log(res)); //{ pos: 0.9981953501701355, neg: 0.0018045296892523766 }
+
+```
+
+### API
+
+- **loadModel()**
+
+  Loads the model files.
+
+  Running in Node.js the first time will download the model assets locally under `/model` directory. The subsequent calls will load the model from the directory.
+
+  Returns the TensorFlow.js model.
+
+- **processInput(text)**
+
+  Processes the input text to the shape and format expected by the model.
+
+  `text` - sentence to be processed. It should be a sentence with a period although this is not necessary.
+
+  Returns a named tensor map that contains:
+  `{'segment_ids_1': Tensor of shape [128],
+    'input_ids_1': Tensor of shape [128],
+    'input_mask_1': Tensor of shape [128]}`
+
+- **runInference(inputFeatures)**
+
+  Runs inference on the named tensor map passed. The output is a tensor that contains softmax of positive and negative percentages.
+
+  `inputFeature` - a named tensor map representation of a text.
+
+  Returns the inference results as a 1D tensor.
+
+- **processOutput(tensor)**
+
+  Transform the inference output to a Json object.
+
+  `tensor` - the model output from running inference.
+
+  Returns an object containing: `{neg: number, pos: number}`
+
+
+- **predict(text)**
+
+  Loads the model, processes the input text, runs inference, processes the inference output, and returns a prediction object. This is a convenience function to avoid having to call each of the functions (`loadModel`, `processInput`, `runInference`, `processOutput`) individually.
+
+  `text` - sentence to be analyzed. It should be a sentence with a period although this is not necessary.
+
+  Returns an object containing: `{neg: number, pos: number}`
+
+- **encode(text)**
+
+  Tokenize the text as token ids using the BERT 32k vocabularies.
+
+  `text` - sentence to be encoded.
+
+  Returns an array of BERT token ids.
+
+- **idsToTokens(ids)**
+
+  Transform the BERT token ids into tokens.
+
+  `ids` - BERT token ids.
+
+  Returns an array of BERT tokens.
+
+- **version**
+
+  Returns the version
+
+## Model
+
+The model assets produced by converting the pre-trained model to the TensorFlow.js format can be found in the `/model` directory after loadModel is called in Node.js.
+
+## Resources
+
+- [MAX Text Sentiment Classifier](https://developer.ibm.com/exchanges/models/all/max-text-sentiment-classifier/)
+
+## License
+
+[Apache-2.0](https://github.com/CODAIT/max-tfjs-models/blob/master/LICENSE)
diff --git a/bert-text-classifier/examples/test.html b/bert-text-classifier/examples/test.html
new file mode 100644
index 0000000..3a15c18
--- /dev/null
+++ b/bert-text-classifier/examples/test.html
@@ -0,0 +1,14 @@
+<meta charset="utf-8"/>
+<html>
+    <head>
+        <title>text classifier</title>
+        <script src="/Users/tedchang/pub_git/max-tfjs-models/bert-text-classifier/dist/src/max.sentimentclass.js"></script>
+        <script>
+        textSentimentClassifier.encode("i like strawberries").then(ids => console.log(ids));
+        textSentimentClassifier.predict("i like strawberries").then(res => console.log(res));
+        </script>
+    </head>
+    <body>
+
+    </body>
+</html>
diff --git a/bert-text-classifier/examples/test.js b/bert-text-classifier/examples/test.js
new file mode 100644
index 0000000..c53dfa1
--- /dev/null
+++ b/bert-text-classifier/examples/test.js
@@ -0,0 +1,2 @@
+const tc = require("../dist/src/max.sentimentclass.cjs.js");
+tc.predict("i like strawberries").then(res=>console.log(res));
diff --git a/bert-text-classifier/package.json b/bert-text-classifier/package.json
new file mode 100644
index 0000000..f2c6545
--- /dev/null
+++ b/bert-text-classifier/package.json
@@ -0,0 +1,82 @@
+{
+  "name": "@codait/text-sentiment-classifier",
+  "version": "0.1.0",
+  "description": "This model is able to detect whether a text fragment leans towards a positive or a negative sentiment.",
+  "main": "dist/src/max.sentimentclass.cjs.js",
+  "module": "dist/src/max.sentimentclass.es.js",
+  "jsdelivr": "dist/src/max.sentimentclass.js",
+  "unpkg": "dist/max.sentimentclass.js",
+  "scripts": {
+    "clean": "rimraf dist && mkdirp dist",
+    "rollup": "rollup --config rollup.config.js",
+    "uglify": "uglifyjs dist/max.sentimentclass.js -mc --source-map --output dist/max.humanpose.min.js",
+    "copy": "ncp model dist/model",
+    "build": "npm run clean && npm run rollup",
+    "test": "standard && npm run build && jasmine test/test.js",
+    "watch:js": "rollup --config rollup.config.js --watch",
+    "watch:ts": "tsc --watch",
+    "dev": "concurrently -c \"bgBlue.bold,bgMagenta.bold\" \"npm:watch:js\" \"npm:watch:ts\""
+  },
+  "files": [
+    "dist",
+    "model",
+    "src"
+  ],
+  "author": "Ted Chang (https://github.com/tedhtchang/)",
+  "license": "Apache-2.0",
+  "keywords": [
+    "bert text sentiment classifier",
+    "model asset exchange",
+    "max",
+    "tensorflow.js",
+    "tensorflowjs",
+    "tensorflow",
+    "tf.js",
+    "tfjs",
+    "machine learning",
+    "Wordpiece tokenizer"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/CODAIT/max-tfjs-models.git"
+  },
+  "bugs": {
+    "url": "https://github.com/CODAIT/max-tfjs-models/issues"
+  },
+  "homepage": "https://github.com/CODAIT/max-tfjs-models/tree/master/bert-text-classifier#readme",
+  "dependencies": {
+    "@tensorflow-models/universal-sentence-encoder": "^1.2.1",
+    "@tensorflow/tfjs": "^1.5.1",
+    "@tensorflow/tfjs-node": "^1.4.0",
+    "@types/express": "^4.16.1",
+    "@types/tar": "^4.0.3",
+    "express": "^4.17.1",
+    "node-fetch": "^2.6.0",
+    "numpy": "0.0.1",
+    "tar": "^5.0.5"
+  },
+  "devDependencies": {
+    "@types/jasmine": "^3.5.0",
+    "@types/rollup-plugin-json": "^3.0.2",
+    "builtin-modules": "^3.1.0",
+    "concurrently": "^4.1.0",
+    "jasmine": "^3.5.0",
+    "jasmine-ts": "^0.3.0",
+    "mkdirp": "^0.5.1",
+    "ncp": "^2.0.0",
+    "rimraf": "^2.6.3",
+    "rollup": "^1.13.1",
+    "rollup-plugin-json": "^4.0.0",
+    "rollup-plugin-node-resolve": "^5.0.1",
+    "rollup-plugin-replace": "^2.2.0",
+    "rollup-plugin-typescript2": "^0.21.1",
+    "ts-node": "^8.5.4",
+    "typescript": "^3.7.3",
+    "unicode-12.1.0": "^0.8.0"
+  },
+  "standard": {
+    "ignore": [
+      "dist"
+    ]
+  }
+}
diff --git a/bert-text-classifier/rollup.config.js b/bert-text-classifier/rollup.config.js
new file mode 100644
index 0000000..f1fbda1
--- /dev/null
+++ b/bert-text-classifier/rollup.config.js
@@ -0,0 +1,78 @@
+import node from 'rollup-plugin-node-resolve';
+import typescript from 'rollup-plugin-typescript2';
+import json from 'rollup-plugin-json';
+import builtins from 'builtin-modules';
+import replace from 'rollup-plugin-replace';
+
+const jsonPlugin = json({
+  include: './package.json',
+  preferConst: true,
+  indent: '  ',
+  compact: true,
+  namedExports: ['version']
+})
+
+export default[
+  {
+    input: 'src/text-sentiment-classifier.ts',
+    output: [
+      {
+        name: 'textSentimentClassifier',
+        file: 'dist/src/max.sentimentclass.js',
+        format: 'iife',
+        sourcemap: true
+      },
+      {
+        name: 'textSentimentClassifier',
+        file: 'dist/src/max.sentimentclass.es.js',
+        format: 'es',
+        sourcemap: true
+      }
+    ],
+    plugins: [
+      typescript({
+        clean: true,
+        tsconfigOverride: {
+          compilerOptions: {
+            module: 'ES2015',
+            noUnusedLocals: false,
+            inlineSourceMap: false
+          }
+        }
+      }),
+      replace({
+        'server/sentimentanalysis': 'js/sentimentanalysis',
+        include: ['src/text-sentiment-classifier.ts']
+      }),
+      jsonPlugin,
+      node(),
+    ],
+    external: builtins
+  },
+  {
+    input: 'src/text-sentiment-classifier.ts',
+    output:[
+      {
+        name: 'textSentimentClassifier',
+        file: 'dist/src/max.sentimentclass.cjs.js',
+        format: 'cjs',
+        sourcemap: true
+      },
+    ],
+    plugins: [
+      typescript({
+        clean: true,
+        tsconfigOverride: {
+          compilerOptions: {
+            module: 'ES2015',
+            noUnusedLocals: false,
+            inlineSourceMap: false
+          }
+        }
+      }),
+      jsonPlugin,
+      node()
+    ],
+    external: builtins
+  }
+]
diff --git a/bert-text-classifier/src/js/sentimentanalysis.ts b/bert-text-classifier/src/js/sentimentanalysis.ts
new file mode 100644
index 0000000..2d7f5b6
--- /dev/null
+++ b/bert-text-classifier/src/js/sentimentanalysis.ts
@@ -0,0 +1,51 @@
+import WordPieceTokenizer from "../tokenization";
+import * as tf from '@tensorflow/tfjs';
+
+const vocabUrl = 'http://s3.us.cloud-object-storage.appdomain.cloud/bert-sentiment-tfjs/model/vocab.json'
+const modelUrl = 'http://s3.us.cloud-object-storage.appdomain.cloud/bert-sentiment-tfjs/model/model.json';
+
+
+export default class SentimentAnalysis {
+  private _model: tf.GraphModel;
+  private _tokenizer: WordPieceTokenizer;
+
+  public get tokenizer(): WordPieceTokenizer {
+    return this._tokenizer;
+  }
+
+  public get model() : tf.GraphModel {
+    return this._model;
+  }
+
+  async init(){
+    if(! this.model) await this.loadModel();
+    if(! this.tokenizer) await this.loadTokenizer();
+  }
+
+  async loadModel(){
+    this._model = await tf.loadGraphModel(
+      modelUrl, {requestInit: {headers: {"origin": "localhost"}}})
+    // console.log(`Model loaded from ${modelUrl}.`);
+  }
+
+  async loadTokenizer(){
+    this._tokenizer = new WordPieceTokenizer(true);
+    await this.tokenizer.init(vocabUrl);
+    // console.log("Tokenizer loaded.")
+  }
+  /**
+ * Classify a text input and return a json object with pos and neg
+ * sentiment percentages
+ */
+  async analyzeText(text: string){
+    return await this.inference(await this.tokenizer.inputFeature(text));
+  }
+
+  async inference(feature: tf.NamedTensorMap){
+    if (! this.model) await this.loadModel();
+    return tf.tidy(() => {
+      let pred: tf.Tensor = this.model.execute({...feature}, 'loss/Softmax') as tf.Tensor;
+      return pred.squeeze([0]);
+    });
+  }
+}
diff --git a/bert-text-classifier/src/server/sentimentanalysis.ts b/bert-text-classifier/src/server/sentimentanalysis.ts
new file mode 100644
index 0000000..0d10970
--- /dev/null
+++ b/bert-text-classifier/src/server/sentimentanalysis.ts
@@ -0,0 +1,62 @@
+import WordPieceTokenizer from "../tokenization";
+import * as tf from '@tensorflow/tfjs';
+import {fetchModel, untar} from './util';
+
+const vocabUrl = 'http://s3.us.cloud-object-storage.appdomain.cloud/bert-sentiment-tfjs/model/vocab.json'
+const modelArch = 'http://s3.us.cloud-object-storage.appdomain.cloud/bert-sentiment-tfjs/model.tgz';
+
+export default class SentimentAnalysis {
+  private _model: tf.GraphModel;
+  private _tokenizer: WordPieceTokenizer;
+
+  public get tokenizer(): WordPieceTokenizer {
+    return this._tokenizer;
+  }
+
+  public get model() : tf.GraphModel {
+    return this._model;
+  }
+
+  async init(){
+    if(! this.model) await this.loadModel();
+    if(! this.tokenizer) await this.loadTokenizer();
+  }
+
+  async loadModel(){
+    const tfn = require('@tensorflow/tfjs-node');
+    const fs = require('fs');
+    const path = require('path');
+    const modelTgz = path.join(`${__dirname}`, '..', '..', '/model/model.tgz');
+    const modelJson = path.join(`${__dirname}`,'..', '..', '/model/model.json');
+    // console.log(modelJson);
+    const modelDir = path.join(`${__dirname}`, '..', '..','/model');
+    if(!fs.existsSync(modelJson)){
+      await fetchModel(modelArch, modelDir);
+      await untar(modelTgz,modelDir);
+    }
+    const fileSystem = require('@tensorflow/tfjs-node/dist/io/file_system');
+    this._model = await tfn.loadGraphModel(fileSystem.fileSystem(modelJson));
+    // console.log(`Model loaded from ${modelJson}.`);
+  }
+
+  async loadTokenizer(){
+    this._tokenizer = new WordPieceTokenizer(true);
+    await this.tokenizer.init(vocabUrl);
+    // console.log("Tokenizer loaded.")
+  }
+  /**
+ * Classify a text input and return a json object with pos and neg
+ * sentiment percentages
+ */
+  async analyzeText(text: string){
+    return await this.inference(await this.tokenizer.inputFeature(text));
+  }
+
+  async inference(feature: tf.NamedTensorMap){
+    if (! this.model) await this.loadModel();
+    return tf.tidy(() => {
+      let pred: tf.Tensor = this.model.execute({...feature}, 'loss/Softmax') as tf.Tensor;
+      return pred.squeeze([0]);
+    });
+  }
+}
diff --git a/bert-text-classifier/src/server/util.ts b/bert-text-classifier/src/server/util.ts
new file mode 100644
index 0000000..674376a
--- /dev/null
+++ b/bert-text-classifier/src/server/util.ts
@@ -0,0 +1,29 @@
+import fetch from 'node-fetch';
+import fs from 'fs';
+import path from 'path';
+
+
+export async function fetchModel(url: string, dir: string){
+  const target = path.join(dir, 'model.tgz');
+  const res = await fetch(url);
+  // console.log("in fetching mode model");
+  const fileStream = fs.createWriteStream(target);
+  return new Promise((resolve, reject) => {
+    res.body.pipe(fileStream);
+    res.body.on("error", (err)=>{
+      reject(err);
+    });
+    fileStream.on("finish", ()=>{
+      resolve();
+    })
+  });
+}
+export async function untar(source: string, dest: string){
+  const {extract} = require('tar');
+  await extract({
+    file: source,
+    cwd: dest,
+    strip: 1
+  })
+}
+
diff --git a/bert-text-classifier/src/test.ts b/bert-text-classifier/src/test.ts
new file mode 100644
index 0000000..036ddf7
--- /dev/null
+++ b/bert-text-classifier/src/test.ts
@@ -0,0 +1,79 @@
+import sa from '../src/text-sentiment-classifier';
+import * as tf from '@tensorflow/tfjs';
+
+const posText = "i like strawberries";
+const negText = "i hate strawberries";
+const posTokenids = [ 1045, 2066, 13137, 20968 ];
+const posTokens = ['▁i', '▁like', '▁straw', 'berries'];
+const resTensor = tf.tensor1d([0.4, 0.6], 'float32');
+
+function genInput(){
+    // Novel input lengths force recompilation which slows down inference, so it's a good idea to enforce a max length.
+    const EXAMPLE_INPUT_LENGTH = 15;
+    const MAX_INPUT_LENGTH = 128;
+    // This is the tokenization of '[CLS] Hello, my dog is cute. [SEP]'.
+    const input_ids = tf.tensor1d([101, 7592, 1010, 2026, 3899, 2003, 10140, 1012, 102, 0, 0, 0, 0, 0, 0], 'int32')
+        .pad([[0,MAX_INPUT_LENGTH - EXAMPLE_INPUT_LENGTH]]).expandDims();
+    const segment_ids = tf.tensor1d(
+        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'int32')
+        .pad([[0, MAX_INPUT_LENGTH - EXAMPLE_INPUT_LENGTH]]).expandDims();
+    const input_mask = tf.tensor1d(
+        [1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0], 'int32')
+        .pad([[0, MAX_INPUT_LENGTH - EXAMPLE_INPUT_LENGTH]]).expandDims(0);
+    return {'segment_ids_1': segment_ids,
+            'input_ids_1': input_ids,
+            'input_mask_1': input_mask}
+}
+
+describe('Bert Sentiment Classifier', () => {
+    it('version returns a valid version number', () => {
+        expect(sa.version).toMatch(/(\d+)\.(\d+)\.(\d+)/);
+    });
+
+    it('encode returns correct the coresponding token ids', async () => {
+        const res = await sa.encode(posText);
+        expect(res).toEqual(posTokenids);
+    });
+
+    it('runInference takes proper named tensor map and returns a tensor', async () => {
+        const res = await sa.runInference(genInput());
+        expect(res).toBeInstanceOf(tf.Tensor);
+    });
+
+    it('idsToTokens convert token ids back to tokens', async () => {
+        const res = await sa.idsToTokens(posTokenids);
+        expect(res).toEqual(posTokens);
+    });
+
+    it('processOutput convert a result Tensor to a Json object', () => {
+        const res =  sa.processOutput(resTensor);
+        expect(res.neg).toBeInstanceOf(Number);
+        expect(res.pos).toBeInstanceOf(Number);
+    })
+
+    it('processInput convert text to input for inference', async () => {
+        const res = await sa.processInput(posText);
+        expect(res.input_ids_1.shape).toEqual([1,128]);
+        expect(res.segment_ids_1.shape).toEqual([1,128]);
+        expect(res.input_mask_1.shape).toEqual([1,128]);
+    });
+    beforeEach(function() {
+    //   var originalTimeout = jasmine.DEFAULT_TIMEOUT_INTERVAL;
+      jasmine.DEFAULT_TIMEOUT_INTERVAL = 60000;
+    });
+
+    it('predict() leans towards pos prediction', async () => {
+        const res = await sa.predict(posText);
+        expect(res.pos).toBeGreaterThan(res.neg);
+      });
+
+    it('predict() leans towards neg prediction', async () => {
+        const res = await sa.predict(negText);
+        expect(res.neg).toBeGreaterThan(res.pos);
+    });
+
+    it('The sum of neg and pos of prediction should be close to 1', async () => {
+        const res = await sa.predict(negText);
+        expect(res.neg + res.pos).toBeCloseTo(1);
+    });
+});
diff --git a/bert-text-classifier/src/text-sentiment-classifier.ts b/bert-text-classifier/src/text-sentiment-classifier.ts
new file mode 100644
index 0000000..dc36a31
--- /dev/null
+++ b/bert-text-classifier/src/text-sentiment-classifier.ts
@@ -0,0 +1,62 @@
+import * as tf from "@tensorflow/tfjs";
+import SentimentAnalysis from "./server/sentimentanalysis";
+import packageJson from "../package.json";
+
+export interface SentRes{
+  neg: number,
+  pos: number
+}
+
+async function processInput(text: string): Promise<tf.NamedTensorMap>{
+  const sa = new SentimentAnalysis();
+  await sa.loadTokenizer()
+  return await sa.tokenizer.inputFeature(text);
+}
+
+async function loadModel(): Promise<tf.GraphModel> {
+  const sa = new SentimentAnalysis();
+  await sa.loadModel()
+  return sa.model;
+}
+
+async function runInference(feature: tf.NamedTensorMap): Promise<tf.Tensor>{
+  const sa = new SentimentAnalysis();
+  await sa.init()
+  return await sa.inference(feature);
+}
+
+function processOutput(res: tf.Tensor): SentRes{
+  const arr = res.arraySync() as number[];
+  return {"pos": arr[0], "neg": arr[1]};
+}
+
+async function predict(text: string): Promise<SentRes>{
+  const sa = new SentimentAnalysis();
+  await sa.init();
+  let res = await sa.analyzeText(text);
+  return processOutput(res);
+}
+
+async function encode(text: string): Promise<number[]>{
+  const sa = new SentimentAnalysis();
+  await sa.loadTokenizer()
+  return await sa.tokenizer.tokenize(text);
+}
+
+async function idsToTokens(ids: number[]){
+  const sa = new SentimentAnalysis();
+  await sa.loadTokenizer();
+  return sa.tokenizer.convertIdsToTokens(ids);
+}
+
+
+export default {
+  processInput,
+  loadModel,
+  runInference,
+  processOutput,
+  predict,
+  encode,
+  idsToTokens,
+  version: packageJson.version
+}
diff --git a/bert-text-classifier/src/tokenization.ts b/bert-text-classifier/src/tokenization.ts
new file mode 100644
index 0000000..9680c81
--- /dev/null
+++ b/bert-text-classifier/src/tokenization.ts
@@ -0,0 +1,184 @@
+import * as use from '@tensorflow-models/universal-sentence-encoder';
+import * as tf from '@tensorflow/tfjs';
+
+function isPunctuation(cp: number):boolean {
+  // Checks a cp is a punctuation character or not.
+  return (punctuations.indexOf(cp) !== -1);
+}
+
+function runStripAccents(text: string){
+  // strips accent marks from text
+  text = text.normalize("NFD");
+  nsmarks.forEach((cp: number) => {text = text.replace(String.fromCodePoint(cp), "")});
+  return text;
+}
+
+function isWhiteSpace(cp: number): boolean{
+  // \t, \n, and \r are technically control characters but we treat them
+  // as whitespace since they are generally considered as such.
+  if (cp === 32 || cp === 9 || cp === 10 || cp === 13) {
+    return true;
+  }
+  if (spcsep.indexOf(cp) != -1){
+    return true;
+  }
+  return false;
+}
+
+function isControl(cp: number): boolean{
+  // "\t" "\n" "\r" are technically control characters but we count them as whitespace
+  // characters.
+  if (ctlchar.indexOf(cp) !== -1 && cp !== 9 && cp !== 10 && cp !== 13){
+    return true;
+  }
+  return false;
+}
+
+function runSplitOnPunctuation(text: string){
+  // Splits punctuation with a space on a piece of text
+  // e.g.:
+  // "abc?" -> "abc ?"
+  // "abc?def" -> "abc ? def"
+  // "abc??def" -> abc ? ? def"
+  const output = [];
+  let preCodePoint = -1;
+  let preCodePointIsPunc = false;
+  for (let i = 0; i < text.length; i++){
+      // current cp is a punctuation AND previous codePoint is not a space
+      // e.g abc?
+      const codePoint = text.charCodeAt(i);
+      const isPuncCurrent = isPunctuation(codePoint);
+      if (isPuncCurrent && preCodePoint !== 32) {
+        output.push(32);
+        // previous cp is not space AND previous codePoint is punctuation
+        // e.g. abc??
+      } else if (preCodePointIsPunc && codePoint !== 32){
+        output.push(32);
+      }
+      output.push(codePoint);
+      preCodePoint = codePoint;
+      preCodePointIsPunc = isPuncCurrent;
+  }
+  return String.fromCodePoint(...output);
+}
+
+function cleanText(text: string){
+  //Performs invalid character removal and whitespace cleanup on text.
+  const output = [];
+  for (let i = 0; i < text.length; i++){
+    let cp = text.charCodeAt(i);
+    if (cp === 0 || cp === 65533 || isControl(cp)){
+      continue;
+    }
+    if (isWhiteSpace(cp)){
+      output.push(32);
+    }
+    else {
+      output.push(cp);
+    }
+  }
+  return String.fromCharCode(...output);
+}
+
+export default class WordPieceTokenizer{
+  //Runs basic tokenization (punctuation splitting, lower casing, etc.).
+  private doLowerCase: boolean;
+  tokenizer: use.Tokenizer;
+  clsId: number;
+  sepId: number;
+
+  constructor(doLowerCase: boolean){
+    this.doLowerCase = doLowerCase;
+  }
+
+  async init(pathToVocabulary: string, ) {
+    await this.loadTokenizer(pathToVocabulary);
+    this.clsId = this.convertTokenToId('[CLS]')[0];
+    this.sepId = this.convertTokenToId('[SEP]')[0];
+  }
+
+  async tokenize(text: string){
+    text = cleanText(text);
+    text = runSplitOnPunctuation(text);
+    text = runStripAccents(text);
+    if (this.doLowerCase){
+      text = text.toLowerCase()
+    }
+    return this.tokenizer.encode(text);
+  }
+  async loadTokenizer(pathToVocabulary: string){
+    this.tokenizer = await use.loadTokenizer(pathToVocabulary);
+    //console.log("Loaded Tokenizer.");
+  }
+
+  convertIdsToTokens(ids: number[]){
+    let tokens: string[] = [];
+    ids.forEach( id => {tokens.push(this.tokenizer.vocabulary[id][0])});
+    return tokens;
+  }
+
+  convertTokenToId(token: string){
+    //convert a token directly to token ID without any pre processing
+    return this.tokenizer.encode(token);
+  }
+
+  async inputFeature(text: string){
+    const singleInput = await this.convertSingleExample(text);
+    let inputIds = tf.tensor1d(singleInput.inputIds, 'int32').expandDims(0);
+    let inputMask = tf.tensor1d(singleInput.inputMask, 'int32').expandDims(0);
+    let segmentIds = tf.tensor1d(singleInput.segmentIds, 'int32').expandDims(0);
+    return {"input_ids_1": inputIds, "input_mask_1": inputMask, "segment_ids_1": segmentIds};
+  }
+
+  async convertSingleExample(text: string){
+    // converts single example to feature input. This is derived from:
+    // https://github.com/google-research/bert/blob/88a817c37f788702a363ff935fd173b6dc6ac0d6/run_classifier.py#L377-L476
+
+    let inputIds: number[] = [];
+    let inputMask: number[] = [];
+    let segmentIds: number[] = [];
+    const tokenIds = await this.tokenize(text);
+    const maxSeqLength = 128;
+
+    inputIds.push(this.clsId)
+    inputMask.push(1);
+    segmentIds.push(0);
+
+    inputIds.push(...tokenIds);
+    tokenIds.forEach(id => {
+      inputMask.push(1);
+      segmentIds.push(0);
+    });
+
+    inputIds.push(this.sepId)
+    inputMask.push(1);
+    segmentIds.push(0);
+
+    // pad with 0 up to the maxSeqLength
+    const numTokens = inputIds.length
+    for (let i = 0; i < maxSeqLength - numTokens; i++){
+      inputIds.push(0);
+      inputMask.push(0);
+      segmentIds.push(0);
+    }
+    // console.log('input_ids: ', inputIds);
+    // console.log('input_mask: ', inputMask);
+    // console.log('segmentIds: ', segmentIds);
+    // console.log('tokens: ', this.convertIdsToTokens(inputIds));
+    return {inputIds, segmentIds, inputMask};
+  }
+}
+
+
+// Unicode code points are extracted from repo below which is intended for node.js not client js.
+// https://github.com/mathiasbynens/unicode-12.1.0/blob/master/General_Category/Nonspacing_Mark/code-points.js
+var nsmarks = [768,769,770,771,772,773,774,775,776,777,778,779,780,781,782,783,784,785,786,787,788,789,790,791,792,793,794,795,796,797,798,799,800,801,802,803,804,805,806,807,808,809,810,811,812,813,814,815,816,817,818,819,820,821,822,823,824,825,826,827,828,829,830,831,832,833,834,835,836,837,838,839,840,841,842,843,844,845,846,847,848,849,850,851,852,853,854,855,856,857,858,859,860,861,862,863,864,865,866,867,868,869,870,871,872,873,874,875,876,877,878,879,1155,1156,1157,1158,1159,1425,1426,1427,1428,1429,1430,1431,1432,1433,1434,1435,1436,1437,1438,1439,1440,1441,1442,1443,1444,1445,1446,1447,1448,1449,1450,1451,1452,1453,1454,1455,1456,1457,1458,1459,1460,1461,1462,1463,1464,1465,1466,1467,1468,1469,1471,1473,1474,1476,1477,1479,1552,1553,1554,1555,1556,1557,1558,1559,1560,1561,1562,1611,1612,1613,1614,1615,1616,1617,1618,1619,1620,1621,1622,1623,1624,1625,1626,1627,1628,1629,1630,1631,1648,1750,1751,1752,1753,1754,1755,1756,1759,1760,1761,1762,1763,1764,1767,1768,1770,1771,1772,1773,1809,1840,1841,1842,1843,1844,1845,1846,1847,1848,1849,1850,1851,1852,1853,1854,1855,1856,1857,1858,1859,1860,1861,1862,1863,1864,1865,1866,1958,1959,1960,1961,1962,1963,1964,1965,1966,1967,1968,2027,2028,2029,2030,2031,2032,2033,2034,2035,2045,2070,2071,2072,2073,2075,2076,2077,2078,2079,2080,2081,2082,2083,2085,2086,2087,2089,2090,2091,2092,2093,2137,2138,2139,2259,2260,2261,2262,2263,2264,2265,2266,2267,2268,2269,2270,2271,2272,2273,2275,2276,2277,2278,2279,2280,2281,2282,2283,2284,2285,2286,2287,2288,2289,2290,2291,2292,2293,2294,2295,2296,2297,2298,2299,2300,2301,2302,2303,2304,2305,2306,2362,2364,2369,2370,2371,2372,2373,2374,2375,2376,2381,2385,2386,2387,2388,2389,2390,2391,2402,2403,2433,2492,2497,2498,2499,2500,2509,2530,2531,2558,2561,2562,2620,2625,2626,2631,2632,2635,2636,2637,2641,2672,2673,2677,2689,2690,2748,2753,2754,2755,2756,2757,2759,2760,2765,2786,2787,2810,2811,2812,2813,2814,2815,2817,2876,2879,2881,2882,2883,2884,2893,2902,2914,2915,2946,3008,3021,3072,3076,3134,3135,3136,3142,3143,3144,3146,3147,3148,3149,3157,3158,3170,3171,3201,3260,3263,3270,3276,3277,3298,3299,3328,3329,3387,3388,3393,3394,3395,3396,3405,3426,3427,3530,3538,3539,3540,3542,3633,3636,3637,3638,3639,3640,3641,3642,3655,3656,3657,3658,3659,3660,3661,3662,3761,3764,3765,3766,3767,3768,3769,3770,3771,3772,3784,3785,3786,3787,3788,3789,3864,3865,3893,3895,3897,3953,3954,3955,3956,3957,3958,3959,3960,3961,3962,3963,3964,3965,3966,3968,3969,3970,3971,3972,3974,3975,3981,3982,3983,3984,3985,3986,3987,3988,3989,3990,3991,3993,3994,3995,3996,3997,3998,3999,4000,4001,4002,4003,4004,4005,4006,4007,4008,4009,4010,4011,4012,4013,4014,4015,4016,4017,4018,4019,4020,4021,4022,4023,4024,4025,4026,4027,4028,4038,4141,4142,4143,4144,4146,4147,4148,4149,4150,4151,4153,4154,4157,4158,4184,4185,4190,4191,4192,4209,4210,4211,4212,4226,4229,4230,4237,4253,4957,4958,4959,5906,5907,5908,5938,5939,5940,5970,5971,6002,6003,6068,6069,6071,6072,6073,6074,6075,6076,6077,6086,6089,6090,6091,6092,6093,6094,6095,6096,6097,6098,6099,6109,6155,6156,6157,6277,6278,6313,6432,6433,6434,6439,6440,6450,6457,6458,6459,6679,6680,6683,6742,6744,6745,6746,6747,6748,6749,6750,6752,6754,6757,6758,6759,6760,6761,6762,6763,6764,6771,6772,6773,6774,6775,6776,6777,6778,6779,6780,6783,6832,6833,6834,6835,6836,6837,6838,6839,6840,6841,6842,6843,6844,6845,6912,6913,6914,6915,6964,6966,6967,6968,6969,6970,6972,6978,7019,7020,7021,7022,7023,7024,7025,7026,7027,7040,7041,7074,7075,7076,7077,7080,7081,7083,7084,7085,7142,7144,7145,7149,7151,7152,7153,7212,7213,7214,7215,7216,7217,7218,7219,7222,7223,7376,7377,7378,7380,7381,7382,7383,7384,7385,7386,7387,7388,7389,7390,7391,7392,7394,7395,7396,7397,7398,7399,7400,7405,7412,7416,7417,7616,7617,7618,7619,7620,7621,7622,7623,7624,7625,7626,7627,7628,7629,7630,7631,7632,7633,7634,7635,7636,7637,7638,7639,7640,7641,7642,7643,7644,7645,7646,7647,7648,7649,7650,7651,7652,7653,7654,7655,7656,7657,7658,7659,7660,7661,7662,7663,7664,7665,7666,7667,7668,7669,7670,7671,7672,7673,7675,7676,7677,7678,7679,8400,8401,8402,8403,8404,8405,8406,8407,8408,8409,8410,8411,8412,8417,8421,8422,8423,8424,8425,8426,8427,8428,8429,8430,8431,8432,11503,11504,11505,11647,11744,11745,11746,11747,11748,11749,11750,11751,11752,11753,11754,11755,11756,11757,11758,11759,11760,11761,11762,11763,11764,11765,11766,11767,11768,11769,11770,11771,11772,11773,11774,11775,12330,12331,12332,12333,12441,12442,42607,42612,42613,42614,42615,42616,42617,42618,42619,42620,42621,42654,42655,42736,42737,43010,43014,43019,43045,43046,43204,43205,43232,43233,43234,43235,43236,43237,43238,43239,43240,43241,43242,43243,43244,43245,43246,43247,43248,43249,43263,43302,43303,43304,43305,43306,43307,43308,43309,43335,43336,43337,43338,43339,43340,43341,43342,43343,43344,43345,43392,43393,43394,43443,43446,43447,43448,43449,43452,43453,43493,43561,43562,43563,43564,43565,43566,43569,43570,43573,43574,43587,43596,43644,43696,43698,43699,43700,43703,43704,43710,43711,43713,43756,43757,43766,44005,44008,44013,64286,65024,65025,65026,65027,65028,65029,65030,65031,65032,65033,65034,65035,65036,65037,65038,65039,65056,65057,65058,65059,65060,65061,65062,65063,65064,65065,65066,65067,65068,65069,65070,65071,66045,66272,66422,66423,66424,66425,66426,68097,68098,68099,68101,68102,68108,68109,68110,68111,68152,68153,68154,68159,68325,68326,68900,68901,68902,68903,69446,69447,69448,69449,69450,69451,69452,69453,69454,69455,69456,69633,69688,69689,69690,69691,69692,69693,69694,69695,69696,69697,69698,69699,69700,69701,69702,69759,69760,69761,69811,69812,69813,69814,69817,69818,69888,69889,69890,69927,69928,69929,69930,69931,69933,69934,69935,69936,69937,69938,69939,69940,70003,70016,70017,70070,70071,70072,70073,70074,70075,70076,70077,70078,70089,70090,70091,70092,70191,70192,70193,70196,70198,70199,70206,70367,70371,70372,70373,70374,70375,70376,70377,70378,70400,70401,70459,70460,70464,70502,70503,70504,70505,70506,70507,70508,70512,70513,70514,70515,70516,70712,70713,70714,70715,70716,70717,70718,70719,70722,70723,70724,70726,70750,70835,70836,70837,70838,70839,70840,70842,70847,70848,70850,70851,71090,71091,71092,71093,71100,71101,71103,71104,71132,71133,71219,71220,71221,71222,71223,71224,71225,71226,71229,71231,71232,71339,71341,71344,71345,71346,71347,71348,71349,71351,71453,71454,71455,71458,71459,71460,71461,71463,71464,71465,71466,71467,71727,71728,71729,71730,71731,71732,71733,71734,71735,71737,71738,72148,72149,72150,72151,72154,72155,72160,72193,72194,72195,72196,72197,72198,72199,72200,72201,72202,72243,72244,72245,72246,72247,72248,72251,72252,72253,72254,72263,72273,72274,72275,72276,72277,72278,72281,72282,72283,72330,72331,72332,72333,72334,72335,72336,72337,72338,72339,72340,72341,72342,72344,72345,72752,72753,72754,72755,72756,72757,72758,72760,72761,72762,72763,72764,72765,72767,72850,72851,72852,72853,72854,72855,72856,72857,72858,72859,72860,72861,72862,72863,72864,72865,72866,72867,72868,72869,72870,72871,72874,72875,72876,72877,72878,72879,72880,72882,72883,72885,72886,73009,73010,73011,73012,73013,73014,73018,73020,73021,73023,73024,73025,73026,73027,73028,73029,73031,73104,73105,73109,73111,73459,73460,92912,92913,92914,92915,92916,92976,92977,92978,92979,92980,92981,92982,94031,94095,94096,94097,94098,113821,113822,119143,119144,119145,119163,119164,119165,119166,119167,119168,119169,119170,119173,119174,119175,119176,119177,119178,119179,119210,119211,119212,119213,119362,119363,119364,121344,121345,121346,121347,121348,121349,121350,121351,121352,121353,121354,121355,121356,121357,121358,121359,121360,121361,121362,121363,121364,121365,121366,121367,121368,121369,121370,121371,121372,121373,121374,121375,121376,121377,121378,121379,121380,121381,121382,121383,121384,121385,121386,121387,121388,121389,121390,121391,121392,121393,121394,121395,121396,121397,121398,121403,121404,121405,121406,121407,121408,121409,121410,121411,121412,121413,121414,121415,121416,121417,121418,121419,121420,121421,121422,121423,121424,121425,121426,121427,121428,121429,121430,121431,121432,121433,121434,121435,121436,121437,121438,121439,121440,121441,121442,121443,121444,121445,121446,121447,121448,121449,121450,121451,121452,121461,121476,121499,121500,121501,121502,121503,121505,121506,121507,121508,121509,121510,121511,121512,121513,121514,121515,121516,121517,121518,121519,122880,122881,122882,122883,122884,122885,122886,122888,122889,122890,122891,122892,122893,122894,122895,122896,122897,122898,122899,122900,122901,122902,122903,122904,122907,122908,122909,122910,122911,122912,122913,122915,122916,122918,122919,122920,122921,122922,123184,123185,123186,123187,123188,123189,123190,123628,123629,123630,123631,125136,125137,125138,125139,125140,125141,125142,125252,125253,125254,125255,125256,125257,125258,917760,917761,917762,917763,917764,917765,917766,917767,917768,917769,917770,917771,917772,917773,917774,917775,917776,917777,917778,917779,917780,917781,917782,917783,917784,917785,917786,917787,917788,917789,917790,917791,917792,917793,917794,917795,917796,917797,917798,917799,917800,917801,917802,917803,917804,917805,917806,917807,917808,917809,917810,917811,917812,917813,917814,917815,917816,917817,917818,917819,917820,917821,917822,917823,917824,917825,917826,917827,917828,917829,917830,917831,917832,917833,917834,917835,917836,917837,917838,917839,917840,917841,917842,917843,917844,917845,917846,917847,917848,917849,917850,917851,917852,917853,917854,917855,917856,917857,917858,917859,917860,917861,917862,917863,917864,917865,917866,917867,917868,917869,917870,917871,917872,917873,917874,917875,917876,917877,917878,917879,917880,917881,917882,917883,917884,917885,917886,917887,917888,917889,917890,917891,917892,917893,917894,917895,917896,917897,917898,917899,917900,917901,917902,917903,917904,917905,917906,917907,917908,917909,917910,917911,917912,917913,917914,917915,917916,917917,917918,917919,917920,917921,917922,917923,917924,917925,917926,917927,917928,917929,917930,917931,917932,917933,917934,917935,917936,917937,917938,917939,917940,917941,917942,917943,917944,917945,917946,917947,917948,917949,917950,917951,917952,917953,917954,917955,917956,917957,917958,917959,917960,917961,917962,917963,917964,917965,917966,917967,917968,917969,917970,917971,917972,917973,917974,917975,917976,917977,917978,917979,917980,917981,917982,917983,917984,917985,917986,917987,917988,917989,917990,917991,917992,917993,917994,917995,917996,917997,917998,917999];
+
+// https://github.com/mathiasbynens/unicode-12.1.0/blob/master/General_Category/Punctuation/code-points.js
+var punctuations = [33,34,35,37,38,39,40,41,42,44,45,46,47,58,59,63,64,91,92,93,95,123,125,161,167,171,182,183,187,191,894,903,1370,1371,1372,1373,1374,1375,1417,1418,1470,1472,1475,1478,1523,1524,1545,1546,1548,1549,1563,1566,1567,1642,1643,1644,1645,1748,1792,1793,1794,1795,1796,1797,1798,1799,1800,1801,1802,1803,1804,1805,2039,2040,2041,2096,2097,2098,2099,2100,2101,2102,2103,2104,2105,2106,2107,2108,2109,2110,2142,2404,2405,2416,2557,2678,2800,3191,3204,3572,3663,3674,3675,3844,3845,3846,3847,3848,3849,3850,3851,3852,3853,3854,3855,3856,3857,3858,3860,3898,3899,3900,3901,3973,4048,4049,4050,4051,4052,4057,4058,4170,4171,4172,4173,4174,4175,4347,4960,4961,4962,4963,4964,4965,4966,4967,4968,5120,5742,5787,5788,5867,5868,5869,5941,5942,6100,6101,6102,6104,6105,6106,6144,6145,6146,6147,6148,6149,6150,6151,6152,6153,6154,6468,6469,6686,6687,6816,6817,6818,6819,6820,6821,6822,6824,6825,6826,6827,6828,6829,7002,7003,7004,7005,7006,7007,7008,7164,7165,7166,7167,7227,7228,7229,7230,7231,7294,7295,7360,7361,7362,7363,7364,7365,7366,7367,7379,8208,8209,8210,8211,8212,8213,8214,8215,8216,8217,8218,8219,8220,8221,8222,8223,8224,8225,8226,8227,8228,8229,8230,8231,8240,8241,8242,8243,8244,8245,8246,8247,8248,8249,8250,8251,8252,8253,8254,8255,8256,8257,8258,8259,8261,8262,8263,8264,8265,8266,8267,8268,8269,8270,8271,8272,8273,8275,8276,8277,8278,8279,8280,8281,8282,8283,8284,8285,8286,8317,8318,8333,8334,8968,8969,8970,8971,9001,9002,10088,10089,10090,10091,10092,10093,10094,10095,10096,10097,10098,10099,10100,10101,10181,10182,10214,10215,10216,10217,10218,10219,10220,10221,10222,10223,10627,10628,10629,10630,10631,10632,10633,10634,10635,10636,10637,10638,10639,10640,10641,10642,10643,10644,10645,10646,10647,10648,10712,10713,10714,10715,10748,10749,11513,11514,11515,11516,11518,11519,11632,11776,11777,11778,11779,11780,11781,11782,11783,11784,11785,11786,11787,11788,11789,11790,11791,11792,11793,11794,11795,11796,11797,11798,11799,11800,11801,11802,11803,11804,11805,11806,11807,11808,11809,11810,11811,11812,11813,11814,11815,11816,11817,11818,11819,11820,11821,11822,11824,11825,11826,11827,11828,11829,11830,11831,11832,11833,11834,11835,11836,11837,11838,11839,11840,11841,11842,11843,11844,11845,11846,11847,11848,11849,11850,11851,11852,11853,11854,11855,12289,12290,12291,12296,12297,12298,12299,12300,12301,12302,12303,12304,12305,12308,12309,12310,12311,12312,12313,12314,12315,12316,12317,12318,12319,12336,12349,12448,12539,42238,42239,42509,42510,42511,42611,42622,42738,42739,42740,42741,42742,42743,43124,43125,43126,43127,43214,43215,43256,43257,43258,43260,43310,43311,43359,43457,43458,43459,43460,43461,43462,43463,43464,43465,43466,43467,43468,43469,43486,43487,43612,43613,43614,43615,43742,43743,43760,43761,44011,64830,64831,65040,65041,65042,65043,65044,65045,65046,65047,65048,65049,65072,65073,65074,65075,65076,65077,65078,65079,65080,65081,65082,65083,65084,65085,65086,65087,65088,65089,65090,65091,65092,65093,65094,65095,65096,65097,65098,65099,65100,65101,65102,65103,65104,65105,65106,65108,65109,65110,65111,65112,65113,65114,65115,65116,65117,65118,65119,65120,65121,65123,65128,65130,65131,65281,65282,65283,65285,65286,65287,65288,65289,65290,65292,65293,65294,65295,65306,65307,65311,65312,65339,65340,65341,65343,65371,65373,65375,65376,65377,65378,65379,65380,65381,65792,65793,65794,66463,66512,66927,67671,67871,67903,68176,68177,68178,68179,68180,68181,68182,68183,68184,68223,68336,68337,68338,68339,68340,68341,68342,68409,68410,68411,68412,68413,68414,68415,68505,68506,68507,68508,69461,69462,69463,69464,69465,69703,69704,69705,69706,69707,69708,69709,69819,69820,69822,69823,69824,69825,69952,69953,69954,69955,70004,70005,70085,70086,70087,70088,70093,70107,70109,70110,70111,70200,70201,70202,70203,70204,70205,70313,70731,70732,70733,70734,70735,70747,70749,70854,71105,71106,71107,71108,71109,71110,71111,71112,71113,71114,71115,71116,71117,71118,71119,71120,71121,71122,71123,71124,71125,71126,71127,71233,71234,71235,71264,71265,71266,71267,71268,71269,71270,71271,71272,71273,71274,71275,71276,71484,71485,71486,71739,72162,72255,72256,72257,72258,72259,72260,72261,72262,72346,72347,72348,72350,72351,72352,72353,72354,72769,72770,72771,72772,72773,72816,72817,73463,73464,73727,74864,74865,74866,74867,74868,92782,92783,92917,92983,92984,92985,92986,92987,92996,93847,93848,93849,93850,94178,113823,121479,121480,121481,121482,121483,125278,125279];
+
+// https://github.com/mathiasbynens/unicode-12.1.0/blob/master/General_Category/Control/code-points.js
+var ctlchar = [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159];
+
+// https://github.com/mathiasbynens/unicode-12.1.0/blob/master/General_Category/Space_Separator/code-points.js
+var spcsep = [32,160,5760,8192,8193,8194,8195,8196,8197,8198,8199,8200,8201,8202,8239,8287,12288];
diff --git a/bert-text-classifier/tsconfig.json b/bert-text-classifier/tsconfig.json
new file mode 100644
index 0000000..42bbfe3
--- /dev/null
+++ b/bert-text-classifier/tsconfig.json
@@ -0,0 +1,27 @@
+{
+  "compilerOptions": {
+    "resolveJsonModule": true,
+    "esModuleInterop": true,
+    "module": "commonjs",
+    "moduleResolution": "node",
+    "noImplicitAny": true,
+    "sourceMap": true,
+    "removeComments": true,
+    "preserveConstEnums": true,
+    "declaration": true,
+    "target": "es5",
+    "lib": ["es2015", "dom"],
+    "outDir": "./dist",
+    "noUnusedLocals": true,
+    "noImplicitReturns": true,
+    "noImplicitThis": true,
+    "alwaysStrict": true,
+    "noUnusedParameters": false,
+    "pretty": true,
+    "noFallthroughCasesInSwitch": true,
+    "allowUnreachableCode": false,
+    "experimentalDecorators": true,
+    "skipLibCheck": true
+  },
+  "include": ["src/**/*"]
+}

From 5c1566023f8f97bfa085e290ef5d1cc16efa8e0f Mon Sep 17 00:00:00 2001
From: Ted Chang <htchang@us.ibm.com>
Date: Tue, 28 Jan 2020 18:16:53 -0800
Subject: [PATCH 2/7] addressed some reviewer comments

---
 bert-text-classifier/README.md                 | 2 +-
 bert-text-classifier/examples/test.html        | 2 +-
 bert-text-classifier/package.json              | 5 ++---
 bert-text-classifier/spec/support/jasmine.json | 3 +++
 bert-text-classifier/{src => test}/test.ts     | 0
 5 files changed, 7 insertions(+), 5 deletions(-)
 create mode 100644 bert-text-classifier/spec/support/jasmine.json
 rename bert-text-classifier/{src => test}/test.ts (100%)

diff --git a/bert-text-classifier/README.md b/bert-text-classifier/README.md
index 637cc64..142d559 100644
--- a/bert-text-classifier/README.md
+++ b/bert-text-classifier/README.md
@@ -1,6 +1,6 @@
 # MAX for TensorFlow.js: Text Sentiment Classifier
 
-This is a TensorFlow.js port of the [MAX Human Pose Estimator](https://developer.ibm.com/exchanges/models/all/max-text-sentiment-classifier/) This model is able to detect whether a text fragment leans towards a positive or a negative sentiment.
+This is a TensorFlow.js port of the [MAX Text Sentiment Classifier](https://developer.ibm.com/exchanges/models/all/max-text-sentiment-classifier/) This model is able to detect whether a text fragment leans towards a positive or a negative sentiment.
 
 ## Install
 
diff --git a/bert-text-classifier/examples/test.html b/bert-text-classifier/examples/test.html
index 3a15c18..d1f3784 100644
--- a/bert-text-classifier/examples/test.html
+++ b/bert-text-classifier/examples/test.html
@@ -2,7 +2,7 @@
 <html>
     <head>
         <title>text classifier</title>
-        <script src="/Users/tedchang/pub_git/max-tfjs-models/bert-text-classifier/dist/src/max.sentimentclass.js"></script>
+        <script src="/dist/src/max.sentimentclass.js"></script>
         <script>
         textSentimentClassifier.encode("i like strawberries").then(ids => console.log(ids));
         textSentimentClassifier.predict("i like strawberries").then(res => console.log(res));
diff --git a/bert-text-classifier/package.json b/bert-text-classifier/package.json
index f2c6545..750090b 100644
--- a/bert-text-classifier/package.json
+++ b/bert-text-classifier/package.json
@@ -9,10 +9,9 @@
   "scripts": {
     "clean": "rimraf dist && mkdirp dist",
     "rollup": "rollup --config rollup.config.js",
-    "uglify": "uglifyjs dist/max.sentimentclass.js -mc --source-map --output dist/max.humanpose.min.js",
     "copy": "ncp model dist/model",
     "build": "npm run clean && npm run rollup",
-    "test": "standard && npm run build && jasmine test/test.js",
+    "test": "npm run build && jasmine-ts",
     "watch:js": "rollup --config rollup.config.js --watch",
     "watch:ts": "tsc --watch",
     "dev": "concurrently -c \"bgBlue.bold,bgMagenta.bold\" \"npm:watch:js\" \"npm:watch:ts\""
@@ -47,7 +46,7 @@
   "dependencies": {
     "@tensorflow-models/universal-sentence-encoder": "^1.2.1",
     "@tensorflow/tfjs": "^1.5.1",
-    "@tensorflow/tfjs-node": "^1.4.0",
+    "@tensorflow/tfjs-node": "^1.5.1",
     "@types/express": "^4.16.1",
     "@types/tar": "^4.0.3",
     "express": "^4.17.1",
diff --git a/bert-text-classifier/spec/support/jasmine.json b/bert-text-classifier/spec/support/jasmine.json
new file mode 100644
index 0000000..b311d5f
--- /dev/null
+++ b/bert-text-classifier/spec/support/jasmine.json
@@ -0,0 +1,3 @@
+{
+    "spec_files":["test/test.ts"]
+}
diff --git a/bert-text-classifier/src/test.ts b/bert-text-classifier/test/test.ts
similarity index 100%
rename from bert-text-classifier/src/test.ts
rename to bert-text-classifier/test/test.ts

From aa332cb0b6c2302f8534c1c2af5cd455ba4b123b Mon Sep 17 00:00:00 2001
From: Ted Chang <htchang@us.ibm.com>
Date: Wed, 12 Feb 2020 14:42:23 -0800
Subject: [PATCH 3/7] changed model downloading link example code and test

---
 bert-text-classifier/examples/test.html       | 52 +++++++++++++++----
 bert-text-classifier/examples/test.js         |  1 +
 .../src/js/sentimentanalysis.ts               |  4 +-
 .../src/server/sentimentanalysis.ts           | 18 +++----
 .../src/text-sentiment-classifier.ts          | 15 ++----
 bert-text-classifier/test/test.ts             |  4 +-
 6 files changed, 59 insertions(+), 35 deletions(-)

diff --git a/bert-text-classifier/examples/test.html b/bert-text-classifier/examples/test.html
index d1f3784..20ac6d6 100644
--- a/bert-text-classifier/examples/test.html
+++ b/bert-text-classifier/examples/test.html
@@ -1,14 +1,44 @@
 <meta charset="utf-8"/>
 <html>
-    <head>
-        <title>text classifier</title>
-        <script src="/dist/src/max.sentimentclass.js"></script>
-        <script>
-        textSentimentClassifier.encode("i like strawberries").then(ids => console.log(ids));
-        textSentimentClassifier.predict("i like strawberries").then(res => console.log(res));
-        </script>
-    </head>
-    <body>
-
-    </body>
+  <head>
+  <title>text classifier</title>
+  </head>
+  <body onload=loadModel()>
+    <h1>Text Sentiment Classifier</h1>
+    <div id="div1">
+      <input type="text" id="text_input" name="text_input" value="Enter a sentence.">
+      <input type="button" id="submit_text" value="Analyze">
+    </div>
+    <p>
+      <label for="status">status:</label>
+      <output name="status" id="status">Loading Model...</output>
+    </p>
+  </body>
+  <script src="file://///wsl$/Ubuntu/home/tedchang/pub_git/max-tfjs-models/bert-text-classifier/dist/src/max.sentimentclass.js"></script>
+  <script>
+    const submitButton = document.getElementById("submit_text");
+    const statusElement = document.getElementById('status');
+    submitButton.addEventListener("click", runPredict);
+    function loadModel(){
+      submitButton.setAttribute('disabled','true')
+      textSentimentClassifier.loadModel().then( () =>{
+        submitButton.removeAttribute('disabled');
+        updateStatus("Model Loaded", append=false);
+      })
+    }
+    function updateStatus(msg, append=true) {
+      statusElement.innerHTML += msg;
+      if(!append)
+        statusElement.innerHTML = msg;
+    }
+    async function runPredict(){
+      const text = document.getElementById("text_input").value;
+      textSentimentClassifier.predict(text).then((res) =>{
+        updateStatus(`
+          <br> ${text}
+          <br>Positive &#128512;: ${res['pos'].toFixed(4)}
+          <br>Negative &#128534;: ${res['neg'].toFixed(4)}<br>`);
+      }
+  )}
+  </script>
 </html>
diff --git a/bert-text-classifier/examples/test.js b/bert-text-classifier/examples/test.js
index c53dfa1..7674c62 100644
--- a/bert-text-classifier/examples/test.js
+++ b/bert-text-classifier/examples/test.js
@@ -1,2 +1,3 @@
 const tc = require("../dist/src/max.sentimentclass.cjs.js");
 tc.predict("i like strawberries").then(res=>console.log(res));
+tc.encode("i like strawberries").then(res=>console.log(res));
\ No newline at end of file
diff --git a/bert-text-classifier/src/js/sentimentanalysis.ts b/bert-text-classifier/src/js/sentimentanalysis.ts
index 2d7f5b6..b54ddb6 100644
--- a/bert-text-classifier/src/js/sentimentanalysis.ts
+++ b/bert-text-classifier/src/js/sentimentanalysis.ts
@@ -1,8 +1,8 @@
 import WordPieceTokenizer from "../tokenization";
 import * as tf from '@tensorflow/tfjs';
 
-const vocabUrl = 'http://s3.us.cloud-object-storage.appdomain.cloud/bert-sentiment-tfjs/model/vocab.json'
-const modelUrl = 'http://s3.us.cloud-object-storage.appdomain.cloud/bert-sentiment-tfjs/model/model.json';
+const vocabUrl = 'https://s3.us-south.cloud-object-storage.appdomain.cloud/max-assets-prod/max-text-sentiment-classifier/tfjs/0.1.0/vocab.json'
+const modelUrl = 'https://s3.us-south.cloud-object-storage.appdomain.cloud/max-assets-prod/max-text-sentiment-classifier/tfjs/0.1.0/model.json';
 
 
 export default class SentimentAnalysis {
diff --git a/bert-text-classifier/src/server/sentimentanalysis.ts b/bert-text-classifier/src/server/sentimentanalysis.ts
index 0d10970..b3a5869 100644
--- a/bert-text-classifier/src/server/sentimentanalysis.ts
+++ b/bert-text-classifier/src/server/sentimentanalysis.ts
@@ -1,9 +1,9 @@
 import WordPieceTokenizer from "../tokenization";
 import * as tf from '@tensorflow/tfjs';
-import {fetchModel, untar} from './util';
+import {IORouter} from '@tensorflow/tfjs-core/dist/io/router_registry';
 
-const vocabUrl = 'http://s3.us.cloud-object-storage.appdomain.cloud/bert-sentiment-tfjs/model/vocab.json'
-const modelArch = 'http://s3.us.cloud-object-storage.appdomain.cloud/bert-sentiment-tfjs/model.tgz';
+const vocabUrl = 'https://s3.us-south.cloud-object-storage.appdomain.cloud/max-assets-prod/max-text-sentiment-classifier/tfjs/0.1.0/vocab.json'
+const modelJsonUrl = 'https://s3.us-south.cloud-object-storage.appdomain.cloud/max-assets-prod/max-text-sentiment-classifier/tfjs/0.1.0/model.json'
 
 export default class SentimentAnalysis {
   private _model: tf.GraphModel;
@@ -26,23 +26,19 @@ export default class SentimentAnalysis {
     const tfn = require('@tensorflow/tfjs-node');
     const fs = require('fs');
     const path = require('path');
-    const modelTgz = path.join(`${__dirname}`, '..', '..', '/model/model.tgz');
     const modelJson = path.join(`${__dirname}`,'..', '..', '/model/model.json');
-    // console.log(modelJson);
     const modelDir = path.join(`${__dirname}`, '..', '..','/model');
     if(!fs.existsSync(modelJson)){
-      await fetchModel(modelArch, modelDir);
-      await untar(modelTgz,modelDir);
+      console.log('Downloading Model...');
+      tf.io.registerLoadRouter(tfn.io.http as IORouter);
+      await tf.io.copyModel(modelJsonUrl, 'file://' + modelDir);
     }
-    const fileSystem = require('@tensorflow/tfjs-node/dist/io/file_system');
-    this._model = await tfn.loadGraphModel(fileSystem.fileSystem(modelJson));
-    // console.log(`Model loaded from ${modelJson}.`);
+    this._model = await tfn.loadGraphModel('file://' + modelJson);
   }
 
   async loadTokenizer(){
     this._tokenizer = new WordPieceTokenizer(true);
     await this.tokenizer.init(vocabUrl);
-    // console.log("Tokenizer loaded.")
   }
   /**
  * Classify a text input and return a json object with pos and neg
diff --git a/bert-text-classifier/src/text-sentiment-classifier.ts b/bert-text-classifier/src/text-sentiment-classifier.ts
index dc36a31..0ca355a 100644
--- a/bert-text-classifier/src/text-sentiment-classifier.ts
+++ b/bert-text-classifier/src/text-sentiment-classifier.ts
@@ -7,20 +7,18 @@ export interface SentRes{
   pos: number
 }
 
+const sa = new SentimentAnalysis();
 async function processInput(text: string): Promise<tf.NamedTensorMap>{
-  const sa = new SentimentAnalysis();
-  await sa.loadTokenizer()
+  if(! sa.tokenizer) await sa.loadTokenizer();
   return await sa.tokenizer.inputFeature(text);
 }
 
 async function loadModel(): Promise<tf.GraphModel> {
-  const sa = new SentimentAnalysis();
-  await sa.loadModel()
+  if(! sa.model) await sa.loadModel();
   return sa.model;
 }
 
 async function runInference(feature: tf.NamedTensorMap): Promise<tf.Tensor>{
-  const sa = new SentimentAnalysis();
   await sa.init()
   return await sa.inference(feature);
 }
@@ -31,21 +29,18 @@ function processOutput(res: tf.Tensor): SentRes{
 }
 
 async function predict(text: string): Promise<SentRes>{
-  const sa = new SentimentAnalysis();
   await sa.init();
   let res = await sa.analyzeText(text);
   return processOutput(res);
 }
 
 async function encode(text: string): Promise<number[]>{
-  const sa = new SentimentAnalysis();
-  await sa.loadTokenizer()
+  if(! sa.tokenizer) await sa.loadTokenizer();
   return await sa.tokenizer.tokenize(text);
 }
 
 async function idsToTokens(ids: number[]){
-  const sa = new SentimentAnalysis();
-  await sa.loadTokenizer();
+  if(! sa.tokenizer) await sa.loadTokenizer();
   return sa.tokenizer.convertIdsToTokens(ids);
 }
 
diff --git a/bert-text-classifier/test/test.ts b/bert-text-classifier/test/test.ts
index 036ddf7..54d5719 100644
--- a/bert-text-classifier/test/test.ts
+++ b/bert-text-classifier/test/test.ts
@@ -1,3 +1,5 @@
+/* globals jasmine, describe, it, expect, tf */
+
 import sa from '../src/text-sentiment-classifier';
 import * as tf from '@tensorflow/tfjs';
 
@@ -59,7 +61,7 @@ describe('Bert Sentiment Classifier', () => {
     });
     beforeEach(function() {
     //   var originalTimeout = jasmine.DEFAULT_TIMEOUT_INTERVAL;
-      jasmine.DEFAULT_TIMEOUT_INTERVAL = 60000;
+      jasmine.DEFAULT_TIMEOUT_INTERVAL = 120000;
     });
 
     it('predict() leans towards pos prediction', async () => {

From 61722f5d1efc49cee2cde33a6250e340ee39d096 Mon Sep 17 00:00:00 2001
From: Ted Chang <htchang@us.ibm.com>
Date: Wed, 12 Feb 2020 14:55:06 -0800
Subject: [PATCH 4/7] add conversion manifest

---
 .../model/conversion_manifest.json            | 124 ++++++++++++++++++
 1 file changed, 124 insertions(+)
 create mode 100644 bert-text-classifier/model/conversion_manifest.json

diff --git a/bert-text-classifier/model/conversion_manifest.json b/bert-text-classifier/model/conversion_manifest.json
new file mode 100644
index 0000000..bf7de50
--- /dev/null
+++ b/bert-text-classifier/model/conversion_manifest.json
@@ -0,0 +1,124 @@
+{
+  "name": "Max-Text-Sentiment-Classifier",
+  "url": "https://developer.ibm.com/exchanges/models/all/max-text-sentiment-classifier/",
+  "source": "max-text-sentiment-classifier/1.0/assets.tar.gz/sentiment_BERT_base_uncased/",
+  "framework": "tf_js",
+  "converter": {
+    "tensorflowjs": {
+      "version": "1.2.9",
+      "params": {
+        "output_node_names": "loss/Softmax",
+        "input_format": "tf_saved_model",
+        "output_json": true
+      }
+    }
+  },
+  "output": [
+    "model.json",
+	"group1-shard1of105.bin",
+	"group1-shard2of105.bin",
+	"group1-shard3of105.bin",
+	"group1-shard4of105.bin",
+	"group1-shard5of105.bin",
+	"group1-shard6of105.bin",
+	"group1-shard7of105.bin",
+	"group1-shard8of105.bin",
+	"group1-shard9of105.bin",
+	"group1-shard10of105.bin",
+	"group1-shard11of105.bin",
+	"group1-shard12of105.bin",
+	"group1-shard13of105.bin",
+	"group1-shard14of105.bin",
+	"group1-shard15of105.bin",
+	"group1-shard16of105.bin",
+	"group1-shard17of105.bin",
+	"group1-shard18of105.bin",
+	"group1-shard19of105.bin",
+	"group1-shard20of105.bin",
+	"group1-shard21of105.bin",
+	"group1-shard22of105.bin",
+	"group1-shard23of105.bin",
+	"group1-shard24of105.bin",
+	"group1-shard25of105.bin",
+	"group1-shard26of105.bin",
+	"group1-shard27of105.bin",
+	"group1-shard28of105.bin",
+	"group1-shard29of105.bin",
+	"group1-shard30of105.bin",
+	"group1-shard31of105.bin",
+	"group1-shard32of105.bin",
+	"group1-shard33of105.bin",
+	"group1-shard34of105.bin",
+	"group1-shard35of105.bin",
+	"group1-shard36of105.bin",
+	"group1-shard37of105.bin",
+	"group1-shard38of105.bin",
+	"group1-shard39of105.bin",
+	"group1-shard40of105.bin",
+	"group1-shard41of105.bin",
+	"group1-shard42of105.bin",
+	"group1-shard43of105.bin",
+	"group1-shard44of105.bin",
+	"group1-shard45of105.bin",
+	"group1-shard46of105.bin",
+	"group1-shard47of105.bin",
+	"group1-shard48of105.bin",
+	"group1-shard49of105.bin",
+	"group1-shard50of105.bin",
+	"group1-shard51of105.bin",
+	"group1-shard52of105.bin",
+	"group1-shard53of105.bin",
+	"group1-shard54of105.bin",
+	"group1-shard55of105.bin",
+	"group1-shard56of105.bin",
+	"group1-shard57of105.bin",
+	"group1-shard58of105.bin",
+	"group1-shard59of105.bin",
+	"group1-shard60of105.bin",
+	"group1-shard61of105.bin",
+	"group1-shard62of105.bin",
+	"group1-shard63of105.bin",
+	"group1-shard64of105.bin",
+	"group1-shard65of105.bin",
+	"group1-shard66of105.bin",
+	"group1-shard67of105.bin",
+	"group1-shard68of105.bin",
+	"group1-shard69of105.bin",
+	"group1-shard70of105.bin",
+	"group1-shard71of105.bin",
+	"group1-shard72of105.bin",
+	"group1-shard73of105.bin",
+	"group1-shard74of105.bin",
+	"group1-shard75of105.bin",
+	"group1-shard76of105.bin",
+	"group1-shard77of105.bin",
+	"group1-shard78of105.bin",
+	"group1-shard79of105.bin",
+	"group1-shard80of105.bin",
+	"group1-shard81of105.bin",
+	"group1-shard82of105.bin",
+	"group1-shard83of105.bin",
+	"group1-shard84of105.bin",
+	"group1-shard85of105.bin",
+	"group1-shard86of105.bin",
+	"group1-shard87of105.bin",
+	"group1-shard88of105.bin",
+	"group1-shard89of105.bin",
+	"group1-shard90of105.bin",
+	"group1-shard91of105.bin",
+	"group1-shard92of105.bin",
+	"group1-shard93of105.bin",
+	"group1-shard94of105.bin",
+	"group1-shard95of105.bin",
+	"group1-shard96of105.bin",
+	"group1-shard97of105.bin",
+	"group1-shard98of105.bin",
+	"group1-shard99of105.bin",
+	"group1-shard100of105.bin",
+	"group1-shard101of105.bin",
+	"group1-shard102of105.bin",
+	"group1-shard103of105.bin",
+	"group1-shard104of105.bin",
+	"group1-shard105of105.bin"
+  ]
+}
\ No newline at end of file

From 5c1647104d2c08fb7510c2ac7627890313c7e77f Mon Sep 17 00:00:00 2001
From: ted chang <htchang@us.ibm.com>
Date: Mon, 17 Feb 2020 12:08:32 -0800
Subject: [PATCH 5/7] change relative path

---
 bert-text-classifier/examples/test.html | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/bert-text-classifier/examples/test.html b/bert-text-classifier/examples/test.html
index 20ac6d6..6d31a8d 100644
--- a/bert-text-classifier/examples/test.html
+++ b/bert-text-classifier/examples/test.html
@@ -14,7 +14,7 @@ <h1>Text Sentiment Classifier</h1>
       <output name="status" id="status">Loading Model...</output>
     </p>
   </body>
-  <script src="file://///wsl$/Ubuntu/home/tedchang/pub_git/max-tfjs-models/bert-text-classifier/dist/src/max.sentimentclass.js"></script>
+  <script src="../dist/src/max.sentimentclass.js"></script>
   <script>
     const submitButton = document.getElementById("submit_text");
     const statusElement = document.getElementById('status');

From 911b8eb943c847cef6a1eb3fd0d58bb29248ad1e Mon Sep 17 00:00:00 2001
From: ted chang <htchang@us.ibm.com>
Date: Tue, 18 Feb 2020 17:01:25 -0800
Subject: [PATCH 6/7] Add prediction status and catch clause

---
 bert-text-classifier/examples/test.html       | 61 +++++++++++++------
 bert-text-classifier/examples/test.js         |  2 +-
 .../src/text-sentiment-classifier.ts          |  8 ++-
 3 files changed, 49 insertions(+), 22 deletions(-)

diff --git a/bert-text-classifier/examples/test.html b/bert-text-classifier/examples/test.html
index 6d31a8d..a2e6377 100644
--- a/bert-text-classifier/examples/test.html
+++ b/bert-text-classifier/examples/test.html
@@ -10,35 +10,60 @@ <h1>Text Sentiment Classifier</h1>
       <input type="button" id="submit_text" value="Analyze">
     </div>
     <p>
-      <label for="status">status:</label>
-      <output name="status" id="status">Loading Model...</output>
+      <label for="status">Status:</label>
+      <label name="status" id="status">Loading Model...</label>
+    </p>
+    <p>
+      <label for="status">Result:</label>
+      <label name="result" id="result"></label>
     </p>
   </body>
   <script src="../dist/src/max.sentimentclass.js"></script>
   <script>
-    const submitButton = document.getElementById("submit_text");
-    const statusElement = document.getElementById('status');
+    let submitButton = document.getElementById("submit_text");
+    let statusElement = document.getElementById("status");
+    let resultElement = document.getElementById("result");
     submitButton.addEventListener("click", runPredict);
+    function enableAnalyzeButton(){
+      submitButton.removeAttribute('disabled');
+    }
+
+    function disableAnalyzeButton(){
+      submitButton.setAttribute('disabled','true');
+    }
+
     function loadModel(){
-      submitButton.setAttribute('disabled','true')
+      submitButton.setAttribute('disabled','true');
       textSentimentClassifier.loadModel().then( () =>{
         submitButton.removeAttribute('disabled');
-        updateStatus("Model Loaded", append=false);
+        updateStatus("Model Loaded");
       })
     }
-    function updateStatus(msg, append=true) {
-      statusElement.innerHTML += msg;
-      if(!append)
-        statusElement.innerHTML = msg;
+    function updateStatus(msg) {
+      statusElement.innerHTML = msg;
+    }
+    function appendResult(msg) {
+      resultElement.innerHTML += msg;
     }
-    async function runPredict(){
+    function runPredict(){
+      updateStatus("Running prediction...");
+      disableAnalyzeButton();
       const text = document.getElementById("text_input").value;
-      textSentimentClassifier.predict(text).then((res) =>{
-        updateStatus(`
-          <br> ${text}
-          <br>Positive &#128512;: ${res['pos'].toFixed(4)}
-          <br>Negative &#128534;: ${res['neg'].toFixed(4)}<br>`);
-      }
-  )}
+      setTimeout(() => { //yield the thread to update UI
+        textSentimentClassifier.predict(text)
+        .then((res) =>{
+          appendResult(`
+            <br> ${text}
+            <br>Positive &#128512;: ${res['pos'].toFixed(4)}
+            <br>Negative &#128534;: ${res['neg'].toFixed(4)}<br>`);
+          updateStatus("Ok");
+          enableAnalyzeButton();
+        })
+        .catch((err) => {
+          updateStatus(err);
+          enableAnalyzeButton();
+        })
+      }, 20);
+    }
   </script>
 </html>
diff --git a/bert-text-classifier/examples/test.js b/bert-text-classifier/examples/test.js
index 7674c62..86c58f0 100644
--- a/bert-text-classifier/examples/test.js
+++ b/bert-text-classifier/examples/test.js
@@ -1,3 +1,3 @@
 const tc = require("../dist/src/max.sentimentclass.cjs.js");
 tc.predict("i like strawberries").then(res=>console.log(res));
-tc.encode("i like strawberries").then(res=>console.log(res));
\ No newline at end of file
+tc.encode("i like strawberries").then(res=>console.log(res));
diff --git a/bert-text-classifier/src/text-sentiment-classifier.ts b/bert-text-classifier/src/text-sentiment-classifier.ts
index 0ca355a..925bc76 100644
--- a/bert-text-classifier/src/text-sentiment-classifier.ts
+++ b/bert-text-classifier/src/text-sentiment-classifier.ts
@@ -19,7 +19,7 @@ async function loadModel(): Promise<tf.GraphModel> {
 }
 
 async function runInference(feature: tf.NamedTensorMap): Promise<tf.Tensor>{
-  await sa.init()
+  await sa.init();
   return await sa.inference(feature);
 }
 
@@ -30,8 +30,10 @@ function processOutput(res: tf.Tensor): SentRes{
 
 async function predict(text: string): Promise<SentRes>{
   await sa.init();
-  let res = await sa.analyzeText(text);
-  return processOutput(res);
+  return sa.analyzeText(text)
+  .catch((err) => console.log(err))
+  .then(processOutput);
+
 }
 
 async function encode(text: string): Promise<number[]>{

From 1ab70b00278fc6f9e10ce399885ab77823b11e73 Mon Sep 17 00:00:00 2001
From: Ted Chang <htchang@us.ibm.com>
Date: Tue, 18 Feb 2020 17:59:58 -0800
Subject: [PATCH 7/7] fix duplicate file handler error

---
 bert-text-classifier/package.json                    | 1 -
 bert-text-classifier/src/server/sentimentanalysis.ts | 8 ++++----
 2 files changed, 4 insertions(+), 5 deletions(-)

diff --git a/bert-text-classifier/package.json b/bert-text-classifier/package.json
index 750090b..1defa6a 100644
--- a/bert-text-classifier/package.json
+++ b/bert-text-classifier/package.json
@@ -45,7 +45,6 @@
   "homepage": "https://github.com/CODAIT/max-tfjs-models/tree/master/bert-text-classifier#readme",
   "dependencies": {
     "@tensorflow-models/universal-sentence-encoder": "^1.2.1",
-    "@tensorflow/tfjs": "^1.5.1",
     "@tensorflow/tfjs-node": "^1.5.1",
     "@types/express": "^4.16.1",
     "@types/tar": "^4.0.3",
diff --git a/bert-text-classifier/src/server/sentimentanalysis.ts b/bert-text-classifier/src/server/sentimentanalysis.ts
index b3a5869..702f798 100644
--- a/bert-text-classifier/src/server/sentimentanalysis.ts
+++ b/bert-text-classifier/src/server/sentimentanalysis.ts
@@ -1,9 +1,10 @@
 import WordPieceTokenizer from "../tokenization";
-import * as tf from '@tensorflow/tfjs';
+import * as tf from '@tensorflow/tfjs-node';
 import {IORouter} from '@tensorflow/tfjs-core/dist/io/router_registry';
 
 const vocabUrl = 'https://s3.us-south.cloud-object-storage.appdomain.cloud/max-assets-prod/max-text-sentiment-classifier/tfjs/0.1.0/vocab.json'
 const modelJsonUrl = 'https://s3.us-south.cloud-object-storage.appdomain.cloud/max-assets-prod/max-text-sentiment-classifier/tfjs/0.1.0/model.json'
+tf.io.registerLoadRouter(tf.io.http as IORouter);
 
 export default class SentimentAnalysis {
   private _model: tf.GraphModel;
@@ -23,17 +24,16 @@ export default class SentimentAnalysis {
   }
 
   async loadModel(){
-    const tfn = require('@tensorflow/tfjs-node');
     const fs = require('fs');
     const path = require('path');
     const modelJson = path.join(`${__dirname}`,'..', '..', '/model/model.json');
     const modelDir = path.join(`${__dirname}`, '..', '..','/model');
     if(!fs.existsSync(modelJson)){
       console.log('Downloading Model...');
-      tf.io.registerLoadRouter(tfn.io.http as IORouter);
       await tf.io.copyModel(modelJsonUrl, 'file://' + modelDir);
     }
-    this._model = await tfn.loadGraphModel('file://' + modelJson);
+    const fileSystem = require('@tensorflow/tfjs-node/dist/io/file_system');
+    this._model = await tf.loadGraphModel(fileSystem.fileSystem(modelJson));
   }
 
   async loadTokenizer(){