html-get 2.15.1 → 2.16.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/package.json +4 -4
- package/src/auto-domains.json +266 -1
- package/src/html.js +5 -3
package/package.json
CHANGED
|
@@ -2,7 +2,7 @@
|
|
|
2
2
|
"name": "html-get",
|
|
3
3
|
"description": "Get the HTML from any website, fine-tuned for correction & speed",
|
|
4
4
|
"homepage": "https://nicedoc.com/microlinkhq/html-get",
|
|
5
|
-
"version": "2.
|
|
5
|
+
"version": "2.16.0",
|
|
6
6
|
"main": "src/index.js",
|
|
7
7
|
"bin": {
|
|
8
8
|
"html-get": "bin/index.js"
|
|
@@ -35,7 +35,7 @@
|
|
|
35
35
|
],
|
|
36
36
|
"dependencies": {
|
|
37
37
|
"@kikobeats/time-span": "~1.0.3",
|
|
38
|
-
"@metascraper/helpers": "~5.
|
|
38
|
+
"@metascraper/helpers": "~5.44.0",
|
|
39
39
|
"cheerio": "~1.0.0-rc.12",
|
|
40
40
|
"css-url-regex": "~4.0.0",
|
|
41
41
|
"debug-logfmt": "~1.2.2",
|
|
@@ -44,11 +44,11 @@
|
|
|
44
44
|
"html-encode": "~2.1.6",
|
|
45
45
|
"html-urls": "~2.4.55",
|
|
46
46
|
"is-html-content": "~1.0.0",
|
|
47
|
+
"localhost-url-regex": "~1.0.11",
|
|
47
48
|
"lodash": "~4.17.21",
|
|
48
49
|
"mri": "~1.2.0",
|
|
49
50
|
"p-cancelable": "~2.1.0",
|
|
50
51
|
"p-retry": "~4.6.0",
|
|
51
|
-
"replace-string": "~3.1.0",
|
|
52
52
|
"tinyspawn": "~1.2.6",
|
|
53
53
|
"top-sites": "~1.1.205"
|
|
54
54
|
},
|
|
@@ -67,7 +67,7 @@
|
|
|
67
67
|
"nano-staged": "latest",
|
|
68
68
|
"npm-check-updates": "latest",
|
|
69
69
|
"pretty": "latest",
|
|
70
|
-
"puppeteer": "
|
|
70
|
+
"puppeteer": "21",
|
|
71
71
|
"regex-iso-date": "latest",
|
|
72
72
|
"simple-git-hooks": "latest",
|
|
73
73
|
"standard": "latest",
|
package/src/auto-domains.json
CHANGED
|
@@ -1 +1,266 @@
|
|
|
1
|
-
[
|
|
1
|
+
[
|
|
2
|
+
[
|
|
3
|
+
[
|
|
4
|
+
"domainWithoutSuffix",
|
|
5
|
+
"google"
|
|
6
|
+
]
|
|
7
|
+
],
|
|
8
|
+
[
|
|
9
|
+
[
|
|
10
|
+
"domainWithoutSuffix",
|
|
11
|
+
"youtube"
|
|
12
|
+
]
|
|
13
|
+
],
|
|
14
|
+
[
|
|
15
|
+
[
|
|
16
|
+
"domainWithoutSuffix",
|
|
17
|
+
"microsoft"
|
|
18
|
+
]
|
|
19
|
+
],
|
|
20
|
+
[
|
|
21
|
+
[
|
|
22
|
+
"domainWithoutSuffix",
|
|
23
|
+
"apple"
|
|
24
|
+
]
|
|
25
|
+
],
|
|
26
|
+
[
|
|
27
|
+
[
|
|
28
|
+
"domainWithoutSuffix",
|
|
29
|
+
"wordpress"
|
|
30
|
+
]
|
|
31
|
+
],
|
|
32
|
+
[
|
|
33
|
+
[
|
|
34
|
+
"domainWithoutSuffix",
|
|
35
|
+
"wikipedia"
|
|
36
|
+
]
|
|
37
|
+
],
|
|
38
|
+
[
|
|
39
|
+
[
|
|
40
|
+
"domainWithoutSuffix",
|
|
41
|
+
"blogspot"
|
|
42
|
+
]
|
|
43
|
+
],
|
|
44
|
+
[
|
|
45
|
+
[
|
|
46
|
+
"domainWithoutSuffix",
|
|
47
|
+
"vimeo"
|
|
48
|
+
]
|
|
49
|
+
],
|
|
50
|
+
[
|
|
51
|
+
[
|
|
52
|
+
"domainWithoutSuffix",
|
|
53
|
+
"github"
|
|
54
|
+
]
|
|
55
|
+
],
|
|
56
|
+
[
|
|
57
|
+
[
|
|
58
|
+
"domainWithoutSuffix",
|
|
59
|
+
"nytimes"
|
|
60
|
+
]
|
|
61
|
+
],
|
|
62
|
+
[
|
|
63
|
+
[
|
|
64
|
+
"domainWithoutSuffix",
|
|
65
|
+
"theguardian"
|
|
66
|
+
]
|
|
67
|
+
],
|
|
68
|
+
[
|
|
69
|
+
[
|
|
70
|
+
"domainWithoutSuffix",
|
|
71
|
+
"imdb"
|
|
72
|
+
]
|
|
73
|
+
],
|
|
74
|
+
[
|
|
75
|
+
[
|
|
76
|
+
"domainWithoutSuffix",
|
|
77
|
+
"bbc"
|
|
78
|
+
]
|
|
79
|
+
],
|
|
80
|
+
[
|
|
81
|
+
[
|
|
82
|
+
"domainWithoutSuffix",
|
|
83
|
+
"slideshare"
|
|
84
|
+
]
|
|
85
|
+
],
|
|
86
|
+
[
|
|
87
|
+
[
|
|
88
|
+
"domainWithoutSuffix",
|
|
89
|
+
"soundcloud"
|
|
90
|
+
]
|
|
91
|
+
],
|
|
92
|
+
[
|
|
93
|
+
[
|
|
94
|
+
"domainWithoutSuffix",
|
|
95
|
+
"huffingtonpost"
|
|
96
|
+
]
|
|
97
|
+
],
|
|
98
|
+
[
|
|
99
|
+
[
|
|
100
|
+
"domainWithoutSuffix",
|
|
101
|
+
"spotify"
|
|
102
|
+
]
|
|
103
|
+
],
|
|
104
|
+
[
|
|
105
|
+
[
|
|
106
|
+
"domainWithoutSuffix",
|
|
107
|
+
"twitter"
|
|
108
|
+
]
|
|
109
|
+
],
|
|
110
|
+
[
|
|
111
|
+
[
|
|
112
|
+
"domainWithoutSuffix",
|
|
113
|
+
"pinterest"
|
|
114
|
+
]
|
|
115
|
+
],
|
|
116
|
+
[
|
|
117
|
+
[
|
|
118
|
+
"domainWithoutSuffix",
|
|
119
|
+
"telegraph"
|
|
120
|
+
]
|
|
121
|
+
],
|
|
122
|
+
[
|
|
123
|
+
[
|
|
124
|
+
"domainWithoutSuffix",
|
|
125
|
+
"techcrunch"
|
|
126
|
+
]
|
|
127
|
+
],
|
|
128
|
+
[
|
|
129
|
+
[
|
|
130
|
+
"domainWithoutSuffix",
|
|
131
|
+
"zoom"
|
|
132
|
+
]
|
|
133
|
+
],
|
|
134
|
+
[
|
|
135
|
+
[
|
|
136
|
+
"domainWithoutSuffix",
|
|
137
|
+
"stackoverflow"
|
|
138
|
+
]
|
|
139
|
+
],
|
|
140
|
+
[
|
|
141
|
+
[
|
|
142
|
+
"domain",
|
|
143
|
+
"abc.net.au"
|
|
144
|
+
]
|
|
145
|
+
],
|
|
146
|
+
[
|
|
147
|
+
[
|
|
148
|
+
"domainWithoutSuffix",
|
|
149
|
+
"yelp"
|
|
150
|
+
]
|
|
151
|
+
],
|
|
152
|
+
[
|
|
153
|
+
[
|
|
154
|
+
"domainWithoutSuffix",
|
|
155
|
+
"eventbrite"
|
|
156
|
+
]
|
|
157
|
+
],
|
|
158
|
+
[
|
|
159
|
+
[
|
|
160
|
+
"domainWithoutSuffix",
|
|
161
|
+
"engadget"
|
|
162
|
+
]
|
|
163
|
+
],
|
|
164
|
+
[
|
|
165
|
+
[
|
|
166
|
+
"domainWithoutSuffix",
|
|
167
|
+
"theverge"
|
|
168
|
+
]
|
|
169
|
+
],
|
|
170
|
+
[
|
|
171
|
+
[
|
|
172
|
+
"domainWithoutSuffix",
|
|
173
|
+
"substack"
|
|
174
|
+
]
|
|
175
|
+
],
|
|
176
|
+
[
|
|
177
|
+
[
|
|
178
|
+
"domainWithoutSuffix",
|
|
179
|
+
"giphy"
|
|
180
|
+
]
|
|
181
|
+
],
|
|
182
|
+
[
|
|
183
|
+
[
|
|
184
|
+
"domainWithoutSuffix",
|
|
185
|
+
"dribbble"
|
|
186
|
+
]
|
|
187
|
+
],
|
|
188
|
+
[
|
|
189
|
+
[
|
|
190
|
+
"domainWithoutSuffix",
|
|
191
|
+
"csdn"
|
|
192
|
+
]
|
|
193
|
+
],
|
|
194
|
+
[
|
|
195
|
+
[
|
|
196
|
+
"domainWithoutSuffix",
|
|
197
|
+
"deviantart"
|
|
198
|
+
]
|
|
199
|
+
],
|
|
200
|
+
[
|
|
201
|
+
[
|
|
202
|
+
"domainWithoutSuffix",
|
|
203
|
+
"digg"
|
|
204
|
+
]
|
|
205
|
+
],
|
|
206
|
+
[
|
|
207
|
+
[
|
|
208
|
+
"domainWithoutSuffix",
|
|
209
|
+
"etsy"
|
|
210
|
+
]
|
|
211
|
+
],
|
|
212
|
+
[
|
|
213
|
+
[
|
|
214
|
+
"domainWithoutSuffix",
|
|
215
|
+
"flickr"
|
|
216
|
+
]
|
|
217
|
+
],
|
|
218
|
+
[
|
|
219
|
+
[
|
|
220
|
+
"domainWithoutSuffix",
|
|
221
|
+
"ghost"
|
|
222
|
+
]
|
|
223
|
+
],
|
|
224
|
+
[
|
|
225
|
+
[
|
|
226
|
+
"domainWithoutSuffix",
|
|
227
|
+
"gitlab"
|
|
228
|
+
]
|
|
229
|
+
],
|
|
230
|
+
[
|
|
231
|
+
[
|
|
232
|
+
"domainWithoutSuffix",
|
|
233
|
+
"imgur"
|
|
234
|
+
]
|
|
235
|
+
],
|
|
236
|
+
[
|
|
237
|
+
[
|
|
238
|
+
"domainWithoutSuffix",
|
|
239
|
+
"meetup"
|
|
240
|
+
]
|
|
241
|
+
],
|
|
242
|
+
[
|
|
243
|
+
[
|
|
244
|
+
"domainWithoutSuffix",
|
|
245
|
+
"producthunt"
|
|
246
|
+
]
|
|
247
|
+
],
|
|
248
|
+
[
|
|
249
|
+
[
|
|
250
|
+
"domainWithoutSuffix",
|
|
251
|
+
"sourceforge"
|
|
252
|
+
]
|
|
253
|
+
],
|
|
254
|
+
[
|
|
255
|
+
[
|
|
256
|
+
"domainWithoutSuffix",
|
|
257
|
+
"tumblr"
|
|
258
|
+
]
|
|
259
|
+
],
|
|
260
|
+
[
|
|
261
|
+
[
|
|
262
|
+
"domainWithoutSuffix",
|
|
263
|
+
"ycombinator"
|
|
264
|
+
]
|
|
265
|
+
]
|
|
266
|
+
]
|
package/src/html.js
CHANGED
|
@@ -1,8 +1,8 @@
|
|
|
1
1
|
'use strict'
|
|
2
2
|
|
|
3
3
|
const { get, split, nth, castArray, forEach } = require('lodash')
|
|
4
|
+
const localhostUrl = require('localhost-url-regex')
|
|
4
5
|
const { TAGS: URL_TAGS } = require('html-urls')
|
|
5
|
-
const replaceString = require('replace-string')
|
|
6
6
|
const isHTML = require('is-html-content')
|
|
7
7
|
const cssUrl = require('css-url-regex')
|
|
8
8
|
const execall = require('execall')
|
|
@@ -95,7 +95,9 @@ const rewriteHtmlUrls = ({ $, url }) => {
|
|
|
95
95
|
const el = $(this)
|
|
96
96
|
const attr = el.attr(urlAttr)
|
|
97
97
|
|
|
98
|
-
if (
|
|
98
|
+
if (localhostUrl().test(attr)) {
|
|
99
|
+
el.remove()
|
|
100
|
+
} else if (typeof attr === 'string' && !attr.startsWith('http')) {
|
|
99
101
|
try {
|
|
100
102
|
const newAttr = new URL(attr, url).toString()
|
|
101
103
|
el.attr(urlAttr, newAttr)
|
|
@@ -117,7 +119,7 @@ const rewriteCssUrls = ({ html, url }) => {
|
|
|
117
119
|
if (cssUrl.startsWith('/')) {
|
|
118
120
|
try {
|
|
119
121
|
const absoluteUrl = new URL(cssUrl, url).toString()
|
|
120
|
-
html =
|
|
122
|
+
html = html.replaceAll(`url(${cssUrl})`, `url(${absoluteUrl})`)
|
|
121
123
|
} catch (_) {}
|
|
122
124
|
}
|
|
123
125
|
})
|