json-web-crawler
v0.8.3
Published
Crawl website by json
Downloads
10
Maintainers
Readme
Json Web Crawler
Use JSON to list all elements (with css 3 and jquery selector) that you want to crawl.
Usage
npm i json-web-crawler --save
const crawl = require('json-web-crawler');
crawl('HTML content', your json setting)
.then(console.log)
.catch(console.error);
Settings
type
The default type is content
.
- content: crawl specific $container to a single json.
- list: crawl a list like Google search result into multi data.
container
DOM element that will focus on. If type is list
, it will crawl each container class.
listOption
Optional, enable in list
type only, use when you don't want to crawl the whole list. ** ALL STRAT FROM 0 **
[ 'limit', 10 ]
: ten elements only (eq(0) ~ eq(9)).[ 'range', 6, 12 ]
: from eq(6) to eq(12 - 1). If without end, it will continue to the last one.[ 'focus', 0, 3, 7, ... ]
: specific elements in list (eq(0), eq(3), eq(7), ...). You can use -1, -2 to count from backward.[ 'ignore', 1, 2, 5 ]
: elements you want to ignore it. You can use -1, -2 to count from backward.
crawl
keyName: { options }
=> keyName: data
crawl: {
image: {
elem: 'img',
get: 'src'
}
}
// will become
image: IMAGE_SRC_URL
options
- elem: element inside
container
. If empty or undefined, it will usecontainer
orlistElems
instead - noChild (boolean): remove all children elem under $(elem)
- outOfContainer (boolean): If exist, It will use $('html').find(elem)
- get: return type of element
text
num
length
: $element.lengthattrName
: $element.attr('attrName')data-dataName
: $element.data('dataNAme')data-dataName:X
:X
is optional.- If data is an array, set
data-dataName:0
will return$elem.data('dataAttribute')[0]
. - If data is an object, set
data-dataName:id
will return$elem.data('dataAttribute')['id']
. - If X not exist, it will return the whole data.
- If data is an array, set
- process: If you want to do something else after 'get' (string type only)
// You can use some simple functions that existed in lodash.
process: [
['match', /regex here/, number], // => str.match(/regex here/)[number], return array if no number, but will cause other process won't work
['split', ',', number], // => str.split(',')[number], return array if no number, but will cause other process won't work
['replace', 'one', 'two'],
['substring', 0, 3],
['prepend', 'text'], // => 'text' + value
['append', 'text'], // => value + 'text'
['indexOf', 'text'] // => return number
['INDENPENDENT_FUNCTION'], // like encodeURI, encodeURIComponent, unescape, etc...
/**
* Due to lodash has the same name `escape` & `unescape` functions with
* different behavior, the origin `escape` & `unescape` function will
* renamed to `encode` & `decode` instead.
*/
],
// Or you want to DIY, you can use function instead
process(value, $elem /* jquery dom */) {
// do something
return newValue;
}
collect: If the value you want is sperated to several elements, use collect to find them all.
- elems: contain multi elements array.
- loop (boolean): It will run all elems (like
li
) you want to get - combineWith: without this, collect will return array
default: return default value when elem not found, null or undefined (
process
will be ignored)
pageNotFound
If match, it will return page not found error.
- elem
- get
- check: like
process
, but only one step
Example
Content Type
Steam Dota2 page in demo
.
const setting = {
type: 'content',
container: '#game_highlights .rightcol',
crawl: {
appId: {
elem: '.glance_tags',
get: 'data-appid'
},
appName: {
outOfContainer: true,
elem: '.apphub_AppName',
get: 'text'
},
image: {
elem: '.game_header_image_full',
get: 'src'
},
reviews: {
elem: '.game_review_summary:eq(0)',
get: 'text',
},
tags: {
elem: '.glance_tags',
collect: {
elems: [{
elem: 'a.app_tag:eq(0)',
get: 'text'
}, {
elem: 'a.app_tag:eq(1)',
get: 'text'
}, {
elem: 'a.app_tag:eq(2)',
get: 'text'
}],
combineWith: ', '
}
},
allTags: {
elem: '.glance_tags a.app_tag',
collect: {
loop: true,
get: 'text',
combineWith: ', '
}
},
description: {
elem: '.game_description_snippet',
get: 'text',
process(value, $elem) {
return value.split(', ');
}
},
releaseDate: {
elem: '.release_date .date',
get: 'text'
}
}
};
List Type
KickStarter popular list in demo
.
const setting = {
pageNotFound: [{
elem: '.grey-frame-inner h1',
get: 'text',
check: ['equal', '404']
}],
type: 'list',
container: '#projects_list .project-card',
listOption: [ 'limit', 3 ],
// listOption: [ 'range', 0, 10 ],
// listOption: [ 'ignore', 0, 2, -1 ],
// listOption: [ 'focus', 3, -3 ],
crawl: {
projectID: {
get: 'data-pid',
},
name: {
elem: '.project-title',
get: 'text',
},
image: {
elem: '.project-thumbnail img',
get: 'src'
},
link: {
elem: '.project-title a',
get: 'href',
process: [
[ 'split', '?', 0 ],
[ 'prepend', 'https://www.kickstarter.com' ]
]
},
description: {
elem: '.project-blurb',
get: 'text'
},
funded: {
elem: '.project-stats-value:eq(0)',
get: 'text'
},
percentPledged: {
elem: '.project-percent-pledged',
get: 'style',
process: [
[ 'split', /:\s?/g, 1 ]
]
},
pledged: {
elem: '.money.usd',
get: 'num'
}
}
};