[Library] Chrome.ahk - Automate Google Chrome using native AutoHotkey. No IE!
Posted: 16 Jan 2018, 17:38
Chrome.ahk
Automate Google Chrome using native AutoHotkey.
Watch the full playlist on YouTube
How it works
Chrome offers a WebSocket based API they call the Chrome DevTools Protocol. This API is what allows web development tools to build integrations, and tools such as Selenium to perform their automation. The protocol's documentation describes a plethora of exciting endpoints accessible using this library, and can be found at the link below.
https://chromedevtools.github.io/devtools-protocol/
Advantages
Limitations
Using this Library
To start using this library you need to create an instance of the class Chrome. Chrome's constructor accepts four optional parameters:
1. ProfilePath - This is the path, relative to the working directory, that your Chrome user profile is located. If an empty folder is given, chrome will generate a new user profile in it. When this parameter is omitted, Chrome will be launched under the default user profile. However, if chrome is already running under that user profile out of debug mode, this will fail. Because of this, it is recommended to always launch Chrome under an alternate user profile.
2. URL - The page that chrome should initially be opened to. Pass an empty string to open Chrome's homepage. When this parameter is omitted, Chrome will be opened to about:blank.
3. ChromePath - The path to find the Chrome executable file. When this parameter is omitted, Chrome will be launched from the path in its start menu entry.
4. DebugPort - The network port to communicate with Chrome over. When this parameter is omitted, port 9222 will be used as specified in the Chrome DevTools Protocol documentation.
Once an instance of the class Chrome has been created, Google Chrome will be launched. To connect to the newly opened page call PageInstance := ChromeInstance.GetPage(). Afterward, use PageInstance.Call() to call protocol endpoints, and PageInstance.Evaluate() to execute JavaScript.
You can find more sample code showing how to use this library in the Examples folder.
Releases
Automate Google Chrome using native AutoHotkey.
Watch the full playlist on YouTube
How it works
Chrome offers a WebSocket based API they call the Chrome DevTools Protocol. This API is what allows web development tools to build integrations, and tools such as Selenium to perform their automation. The protocol's documentation describes a plethora of exciting endpoints accessible using this library, and can be found at the link below.
https://chromedevtools.github.io/devtools-protocol/
Advantages
- No external dependencies such as Selenium are required
- Chrome can be automated even when running in headless mode
- Chrome consistently benchmarks better than Internet Explorer
- Chrome offers extensions which provide unique opportunities for interaction
- Automate your Chromecast
- Connect to remote servers with FoxyProxy and update web based configs
- Manage your password vault with LastPass
- Many features are available that would be difficult to replicate in Internet Explorer
- Page.printToPDF
- Page.captureScreenshot
- Geolocation spoofing
Limitations
- Chrome must be started in debug mode
- If chrome is already running out of debug mode, it must either be closed and reopened or launched again under a new profile that isn't already running
- You cannot attach to an existing non-debug session
- Less flexible than Internet Explorer's COM interface
- Cannot pass function references for callbacks
Using this Library
To start using this library you need to create an instance of the class Chrome. Chrome's constructor accepts four optional parameters:
1. ProfilePath - This is the path, relative to the working directory, that your Chrome user profile is located. If an empty folder is given, chrome will generate a new user profile in it. When this parameter is omitted, Chrome will be launched under the default user profile. However, if chrome is already running under that user profile out of debug mode, this will fail. Because of this, it is recommended to always launch Chrome under an alternate user profile.
2. URL - The page that chrome should initially be opened to. Pass an empty string to open Chrome's homepage. When this parameter is omitted, Chrome will be opened to about:blank.
3. ChromePath - The path to find the Chrome executable file. When this parameter is omitted, Chrome will be launched from the path in its start menu entry.
4. DebugPort - The network port to communicate with Chrome over. When this parameter is omitted, port 9222 will be used as specified in the Chrome DevTools Protocol documentation.
Once an instance of the class Chrome has been created, Google Chrome will be launched. To connect to the newly opened page call PageInstance := ChromeInstance.GetPage(). Afterward, use PageInstance.Call() to call protocol endpoints, and PageInstance.Evaluate() to execute JavaScript.
Code: Select all
#Include Chrome.ahk
; Create an instance of the Chrome class using
; the folder ChromeProfile to store the user profile
FileCreateDir, ChromeProfile
ChromeInst := new Chrome("ChromeProfile")
; Connect to the newly opened tab and navigate to another website
; Note: If your first action is to navigate away, it may be just as
; effective to provide the target URL when instantiating the Chrome class
PageInstance := ChromeInst.GetPage()
PageInstance.Call("Page.navigate", {"url": "https://autohotkey.com/"})
PageInstance.WaitForLoad()
; Execute some JavaScript
PageInstance.Evaluate("alert('Hello World!');")
; Close the browser (note: this closes *all* pages/tabs)
PageInstance.Call("Browser.close")
PageInstance.Disconnect()
ExitApp
returnReleases
I'm very excited to play with this! Thanks GeekDude!
Thanks GeekDude! I'd tried that however I'd failed to close-out the background /default version of Chrome that stays running. After closing that version out, it was able to connect! I'll work on cleaning up the examples I've created and will share tomorrow/soon.
