Chromy is a library for operating headless chrome.
Document Site: https://onetapinc.github.io/chromy/
Chromy is similar to Nightmare.js but has some differences:
- Controlling Chrome via Chrome DevTools Protocol.
- Supports mobile emulation.
- No need to prepare a screen such as xvfb.
- Node 6 or later
- Install Chrome60 or later to your machine before use Chromy.
headless mode is supported by Chrome59 or later.
npm i chromyconst Chromy = require('chromy')
// not headless
// let chromy = new Chromy({visible:true})
let chromy = new Chromy()
chromy.chain()
.goto('http://example.com/')
.evaluate(() => {
return document.querySelectorAll('*').length
})
.result((r) => console.log(r))
.end()
.then(() => chromy.close())You can also use async/await interfaces like this:
const Chromy = require('chromy')
async function main () {
let chromy = new Chromy()
await chromy.goto('http://example.com/')
const result = await chromy.evaluate(() => {
return document.querySelectorAll('*').length
})
console.log(result)
await chromy.close()
}
main()Chromy provides mobile emulation.
The emulation changes a screen resolution, density, userAgent and provides touch emulation.
const Chromy = require('chromy')
let chromy = new Chromy()
chromy.chain()
.emulate('iPhone6')
.goto('http://example.com/')
.tap(100, 100) // emulate tap action by synthesizing touch events.
.evaluate(() => {
return navigator.userAgent
})
.result(console.log)
.end()
.then(() => chromy.close())- Chromy(options)
- .start(startingUrl = null)
- .goto(url, options = {})
- .waitLoadEvent()
- .userAgent(ua)
- Chromy.addCustomDevice(device)
- .emulate(deviceName)
- .forward()
- .back()
- .inject(type, file)
- .evaluate(func|source)
- .result(func)
- .end()
- .exists(selector)
- .visible(selector)
- .wait(msec)
- .wait(selector)
- .wait(func)
- .sleep(msec)
- .type(selector, text)
- .insert(selector, text)
- .check(selector)
- .uncheck(selector)
- .select(selector, value)
- .setFile(selector, files)
- .click(selector, options)
- .mouseMoved(x, y, options = {})
- .mousePressed(x, y, options = {})
- .mouseReleased(x, y, options = {})
- .tap(x, y, options = {})
- .doubleTap(x, y, options = {})
- .scroll(x, y)
- .scrollTo(x, y)
- .rect(selector)
- .rectAll(selector)
- .defineFunction(func)
- .on(eventName, listener)
- .once(eventName, listener)
- .removeListener(eventName, listener)
- .removeAllListeners(eventName)
- .screenshot(options= {})
- .screenshotSelector(selector, options={})
- .screenshotMultipleSelectors(selectors, callback, options = {})
- .screenshotDocument(options = {})
- .pdf(options={})
- .startScreencast(callback, options = {})
- .stopScreencast()
- .console(func)
- .receiveMessage(func)
- .ignoreCertificateErrors()
- .blockUrls(urls)
- .clearBrowserCache()
- .setCookie(params)
- .getCookies(params)
- .deleteCookie(name, url = null)
- .clearAllCookies()
- .clearDataForOrigin (origin = null, type = 'all')
- .getDOMCounters()
- .static cleanup()
- host(default: localhost): host address
- port(default: 9222): --remote-debugging-port
- userDataDir(default: null): Chrome profile path. This option can be used to persist an user profile.
- launchBrowser(default: true): If you want chromy to attach to the Chrome instance that is already launched, set to false.
- visible(default: false): If set to true, chrome is launched in visible mode. This option is not used if launchBrowser is false.
- chromePath(default: null): This option is used to find out an executable of Chrome. If set to null, executable is selected automatically. This option is not used if launchBrowser is false.
- enableExtensions(default: false): Enable extension loading. (Generally, this options is used with userDataDir option)
- chromeFlags(default: []): These flags is passed to Chrome. Each flag must have a prefix string "--". This option is not used if launchBrowser is false.
- waitTimeout(default: 30000): If wait() doesn't finish in the specified time WaitTimeoutError will be thrown.
- gotoTimeout(default: 30000): If goto() doesn't finish in the specified time GotoTimeoutError will be thrown.
- evaluateTimeout(default: 30000): If evaluate() doesn't finish in the specified time EvaluateTimeError will be thrown.
- waitFunctionPollingInterval(default: 100): polling interval for wait().
- typeInterval(default: 20): This option is used only in type() method.
- activateOnStartUp(default: true): activate a first tab on startup. this option is enable only in visible mode.
Launches Chrome browser.
startingUrl: a staring url. If you set to null 'about:blank' is used as a starting url.
Goes to url. If you have not called start(), this method calls start(url) automatically.
waitLoadEvent(default: true): If set to false, goto() doesn't wait until load event is fired.
Returns Response object
wait until a load event is fired.
set a useragent.
ua: new user agent.
add custom device definitions to emulate it.
See src.
emulate a device that is defined by Chromy.addCustomDevice().
go forward to the next page and wait until load event is fired.
go back to the previous page and wait until load event is fired.
Injects a file into browser as a javascript or a css.
type: must be 'js' or 'css' file: injected file.
Evaluates a expression in the browser context.
If the expression returns a Promise object, the promise is resolved automatically.
result() receives a result of previous directive.
chromy.chain()
.goto('http://example.com')
.evaluate(() => {
return document.querySelectorAll('*').length
})
.result((length) => {
// length is a result of evaluate() directive.
console.log(length)
}
.end()Returns whether an node matched with the selector is exists.
Returns whether an node matched with the selector is exists and visible.
alias for .sleep(msec)
wait until selector you specified appear in a DOM tree.
wait until function you supplied is evaluated as true. func() executes in browser window context.
wait for milli seconds you specified.
Sets the files to a file field that matches the selector.
- selector: selector for specifying the file field.
- files: The array or string value that represents a local file path.
waitLoadEvent(default: false): If set to true, wait until load event is fired after click event is fired.
Dispatch mousemoved event.
Dispatch mousedown event.
Dispatch mouseup event.
Synthesize tap by dispatching touch events. (NOTE: To dispatch touch events you need to enable a mobile emulation before.)
Synthesize double tap by dispatching touch events. (NOTE: To dispatch touch events you need to enable a mobile emulation before.)
Scrolls to the position. x and y means relative position.
Scrolls to the position. x and y means absolute position.
Returns a rect of the element specified by selector.
Returns an array of rects that is specified by selector.
function outerFunc () {
return 'VALUE'
}
chromy.chain()
.goto('http://example.com')
.defineFunction(outerFunc)
.evaluate(() => {
outerFunc()
})
.end()Calls DevTools protocol directly.
Adds the listener function.
Adds one time listener function.
Removes the listener function.
Removes all listener function.
Exports a current screen as an image data.
See examples: examples/screenshot.js
- format(default: 'png'): must be either 'png' or 'jpeg'
- quality(default: 100): quality of image.
- fromSurface(default: true): if set to true, take screenshot from surface.
- useDeviceResolution(default: false): if set to true, the image will have same resolution with device.
Exports an area of selector you specified as an image data.
See examples: examples/screenshot.js
Note:
- The size of target specified by selector must be smaller than viewport size. If not, image gets cropped.
- It has a side-effect. After this api is called, scroll position is moved to target position.
See screenshot()
Takes multiple screenshot specified by selector at once. Each image can be received by callback.
Limitation:
- It is impossible that taking a screenshot of the element positioned at below of 16384px because of limitation of chrome. Detail: https://groups.google.com/a/chromium.org/d/msg/headless-dev/DqaAEXyzvR0/P9zmTLMvDQAJ
- selectors: An array of selector
- callback: function(error, image, index, selectors, subIndex)
- error: error information.
- image: image data
- index: index of selectors.
- subIndex: this value is used only if useQuerySelecotrAll is true.
- options:
- model: see explanation of screenDocument()
- format: see explanation of screenshot()
- quality: see explanation of screenshot()
- fromSurface: see explanation of screenshot()
- useQuerySelectorAll(default: false): If set to true, take all the screenshot of elements returned from document.querySelectorAll() (Since v 0.2.13)
Exports a entire document as an image data.
See examples: examples/screenshot.js
Limitation:
- Cannot take a screenshot of an area under 16384px. Detail: https://groups.google.com/a/chromium.org/d/msg/headless-dev/DqaAEXyzvR0/P9zmTLMvDQAJ
Known Issue:
- When this api is called to take large page sometimes strange white area is appeared. This result is caused by --disable-flag option passed to Chrome. After chrome 60 is officially released I remove --disable-flag option to fix this problem.
- model: this parameter affect page size. must be which one of: 'box', 'scroll'. 'box' means box model of body element. 'scroll' means size of scroll area.
- format: see explanation of screenshot()
- quality: see explanation of screenshot()
- fromSurface: see explanation of screenshot()
Exports a current page's printing image as a PDF data. This function is supported only in headless mode (since Chrome60).
See examples: examples/screenshot.js
- options: See devtools protocol
Starts screencast to take screenshots by every frame.
See examples: examples/screencast.js
callback: callback function for receiving parameters of screencastFrame event. See details here options: See details here.
Stops screencast.
chromy.chain()
.goto('http://example.com')
.console((text) => {
console.log(text)
})
.evaluate(() => {
console.log('HEY')
})
.end()receive a message from browser.
You can communicate with a browser by using receiveMessage() and sendToChromy(). sendToChromy() is a special function to communicate with Chromy. When you call receiveMessage() at the first time, sendToChromy() is defined in a browser automatically. A listener function passed to receiveMessage() receives parameters when sendToChromy() is executed in a browser.
chromy.chain()
.goto('http://example.com')
.receiveMessage((msg) => {
console.log(msg[0].value)
})
.evaluate(() => {
sendToChromy({value: 'foo'})
})Ignores all certificate errors.
chromy.chain()
.ignoreCertificateErrors()
.goto('https://xxxxx/')
.end()blocks urls from loading.
urls: array[string]
Wildcard('*') is allowed in url string.
Removes all browser caches.
params: object or array
See chrome document If url parameter is not set, current url(location.href) is used as default value.
name: string or array of string
See chrome document
Remove a cookie.
name: string or array of string url: url associated with cookie. If url is not set, current url(location.href) is used as default value.
Removes all browser cookies.
Clear data for origin.(cookies, local_storage, indexedDb, etc...)
See details here.
Get count of these item: document, node, jsEventListeners
See details here.
close all browsers.
process.on('SIGINT', async () => {
await Chromy.cleanup()
process.exit(1)
})Bug reports and pull requests are welcome on GitHub at https://github.com/OnetapInc/chromy
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent
