forked from tensorflow/tfjs-examples
-
Notifications
You must be signed in to change notification settings - Fork 0
/
data.js
262 lines (239 loc) · 7.69 KB
/
data.js
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
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
/**
* @license
* Copyright 2018 Google LLC. All Rights Reserved.
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
* =============================================================================
*/
import * as https from 'https';
import * as tf from '@tensorflow/tfjs';
// TODO(cais): Support user-supplied text data.
export const TEXT_DATA_URLS = {
'nietzsche': {
url:
'https://storage.googleapis.com/tfjs-examples/lstm-text-generation/data/nietzsche.txt',
needle: 'Nietzsche'
},
'julesverne': {
url:
'https://storage.googleapis.com/tfjs-examples/lstm-text-generation/data/t1.verne.txt',
needle: 'Jules Verne'
},
'shakespeare': {
url:
'https://storage.googleapis.com/tfjs-examples/lstm-text-generation/data/t8.shakespeare.txt',
needle: 'Shakespeare'
},
'tfjs-code': {
url: 'https://cdn.jsdelivr.net/npm/@tensorflow/[email protected]/dist/tf.js',
needle: 'TensorFlow.js Code (Compiled, 0.11.7)'
}
}
/**
* A class for text data.
*
* This class manages the following:
*
* - Converting training data (as a string) into one-hot encoded vectors.
* - Drawing random slices from the training data. This is useful for training
* models and obtaining the seed text for model-based text generation.
*/
export class TextData {
/**
* Constructor of TextData.
*
* @param {string} dataIdentifier An identifier for this instance of TextData.
* @param {string} textString The training text data.
* @param {number} sampleLen Length of each training example, i.e., the input
* sequence length expected by the LSTM model.
* @param {number} sampleStep How many characters to skip when going from one
* example of the training data (in `textString`) to the next.
*/
constructor(dataIdentifier, textString, sampleLen, sampleStep) {
tf.util.assert(
sampleLen > 0,
`Expected sampleLen to be a positive integer, but got ${sampleLen}`);
tf.util.assert(
sampleStep > 0,
`Expected sampleStep to be a positive integer, but got ${sampleStep}`);
if (!dataIdentifier) {
throw new Error('Model identifier is not provided.');
}
this.dataIdentifier_ = dataIdentifier;
this.textString_ = textString;
this.textLen_ = textString.length;
this.sampleLen_ = sampleLen;
this.sampleStep_ = sampleStep;
this.getCharSet_();
this.convertAllTextToIndices_();
}
/**
* Get data identifier.
*
* @returns {string} The data identifier.
*/
dataIdentifier() {
return this.dataIdentifier_;
}
/**
* Get length of the training text data.
*
* @returns {number} Length of training text data.
*/
textLen() {
return this.textLen_;
}
/**
* Get the length of each training example.
*/
sampleLen() {
return this.sampleLen_;
}
/**
* Get the size of the character set.
*
* @returns {number} Size of the character set, i.e., how many unique
* characters there are in the training text data.
*/
charSetSize() {
return this.charSetSize_;
}
/**
* Generate the next epoch of data for training models.
*
* @param {number} numExamples Number examples to generate.
* @returns {[tf.Tensor, tf.Tensor]} `xs` and `ys` Tensors.
* `xs` has the shape of `[numExamples, this.sampleLen, this.charSetSize]`.
* `ys` has the shape of `[numExamples, this.charSetSize]`.
*/
nextDataEpoch(numExamples) {
this.generateExampleBeginIndices_();
if (numExamples == null) {
numExamples = this.exampleBeginIndices_.length;
}
const xsBuffer = new tf.TensorBuffer([
numExamples, this.sampleLen_, this.charSetSize_]);
const ysBuffer = new tf.TensorBuffer([numExamples, this.charSetSize_]);
for (let i = 0; i < numExamples; ++i) {
const beginIndex = this.exampleBeginIndices_[
this.examplePosition_ % this.exampleBeginIndices_.length];
for (let j = 0; j < this.sampleLen_; ++j) {
xsBuffer.set(1, i, j, this.indices_[beginIndex + j]);
}
ysBuffer.set(1, i, this.indices_[beginIndex + this.sampleLen_]);
this.examplePosition_++;
}
return [xsBuffer.toTensor(), ysBuffer.toTensor()];
}
/**
* Get the unique character at given index from the character set.
*
* @param {number} index
* @returns {string} The unique character at `index` of the character set.
*/
getFromCharSet(index) {
return this.charSet_[index];
}
/**
* Convert text string to integer indices.
*
* @param {string} text Input text.
* @returns {number[]} Indices of the characters of `text`.
*/
textToIndices(text) {
const indices = [];
for (let i = 0; i < text.length; ++i) {
indices.push(this.charSet_.indexOf(text[i]));
}
return indices;
}
/**
* Get a random slice of text data.
*
* @returns {[string, number[]} The string and index representation of the
* same slice.
*/
getRandomSlice() {
const startIndex =
Math.round(Math.random() * (this.textLen_ - this.sampleLen_ - 1));
const textSlice = this.slice_(startIndex, startIndex + this.sampleLen_);
return [textSlice, this.textToIndices(textSlice)];
}
/**
* Get a slice of the training text data.
*
* @param {number} startIndex
* @param {number} endIndex
* @param {bool} useIndices Whether to return the indices instead of string.
* @returns {string | Uint16Array} The result of the slicing.
*/
slice_(startIndex, endIndex) {
return this.textString_.slice(startIndex, endIndex);
}
/**
* Get the set of unique characters from text.
*/
getCharSet_() {
this.charSet_ = [];
for (let i = 0; i < this.textLen_; ++i) {
if (this.charSet_.indexOf(this.textString_[i]) === -1) {
this.charSet_.push(this.textString_[i]);
}
}
this.charSetSize_ = this.charSet_.length;
}
/**
* Convert all training text to integer indices.
*/
convertAllTextToIndices_() {
this.indices_ = new Uint16Array(this.textToIndices(this.textString_));
}
/**
* Generate the example-begin indices; shuffle them randomly.
*/
generateExampleBeginIndices_() {
// Prepare beginning indices of examples.
this.exampleBeginIndices_ = [];
for (let i = 0;
i < this.textLen_ - this.sampleLen_ - 1;
i += this.sampleStep_) {
this.exampleBeginIndices_.push(i);
}
// Randomly shuffle the beginning indices.
tf.util.shuffle(this.exampleBeginIndices_);
this.examplePosition_ = 0;
}
}
/**
* Get a file by downloading it if necessary.
*
* @param {string} sourceURL URL to download the file from.
* @param {string} destPath Destination file path on local filesystem.
*/
export async function maybeDownload(sourceURL, destPath) {
const fs = require('fs');
return new Promise(async (resolve, reject) => {
if (!fs.existsSync(destPath) || fs.lstatSync(destPath).size === 0) {
const localZipFile = fs.createWriteStream(destPath);
console.log(`Downloading file from ${sourceURL} to ${destPath}...`);
https.get(sourceURL, response => {
response.pipe(localZipFile);
localZipFile.on('finish', () => {
localZipFile.close(() => resolve());
});
localZipFile.on('error', err => reject(err));
});
} else {
return resolve();
}
});
}