asyncio driver + HTTP server for Chrome devtools protocol
Chrome browser as an HTTP service with an splash compatible HTTP API
Chromewhip is an easily deployable service that runs headless Chrome process wrapped with an HTTP API. Inspired by the `splash <https://github.com/scrapinghub/splash>`__ project, we aim to provide a drop-in replacement for the splash service by adhering to their documented API.
It is currently in early alpha and still being heavily developed. Please use the issue tracker to track the progress towards beta. For now, the required milestone can be summarised as implementing the entire Splash API.
Python 3.6 asyncio driver for Chrome devtools protocol
Chromewhip communicates with the Chrome process with our own asyncio driver.
- Typed Python bindings for devtools protocol through templated generation - get autocomplete with your code editor.
- Can bind events to concurrent commands, which is required for providing a robust HTTP service.
Some example code on how to use it:
import asyncio import logging from chromewhip import Chrome from chromewhip.protocol import page, dom # see logging from chromewhip logging.basicConfig(level=logging.DEBUG) HOST = '127.0.0.1' PORT = 9222 loop = asyncio.get_event_loop() c = Chrome(host=HOST, port=PORT) loop.run_until_complete(c.connect()) tab = c.tabs loop.run_until_complete(tab.enable_page_events()) cmd = page.Page.navigate(url='http://nzherald.co.nz') # send_command will return once the frameStoppedLoading event is received THAT matches # the frameId that it is in the returned command payload. await_on_event_type = page.FrameStoppedLoadingEvent result = loop.run_until_complete(tab.send_command(cmd, await_on_event_type=await_on_event_type)) # send_command always returns a dict with keys `ack` and `event` # `ack` contains the payload on response of a command # `event` contains the payload of the awaited event if `await_on_event_type` is provided ack = result['ack']['result'] event = result['event'] assert ack['frameId'] == event.frameId cmd = page.Page.setDeviceMetricsOverride(width=800, height=600, deviceScaleFactor=0.0, mobile=False, fitWindow=False) loop.run_until_complete(tab.send_command(cmd)) result = loop.run_until_complete(tab.send_command(dom.DOM.getDocument())) dom_obj = result['ack']['result']['root'] # Python types are determined by the `types` fields in the JSON reference for the # devtools protocol, and `send_command` will convert if possible. assert isinstance(dom_obj, dom.Node) print(dom_obj.nodeId) print(dom_obj.nodeName)
Deploying with Docker
docker run --init -it --rm --shm-size=1024m -p=127.0.0.1:8080:8080 --cap-add=SYS_ADMIN \ chuckus/chromewhip
Requirements for MacOS 10.12+
- Google Chrome Canary
Implemented HTTP API
url : string : required
The url to render (required)
js_source : string : optional
viewport : string : optional
View width and height (in pixels) of the browser viewport to render the web page. Format is “x”, e.g. 800x600. Default value is 1024x768.
Query params (including render.html):
- render_all : int : optional
- Possible values are 1 and 0. When render_all=1, extend the viewport to include the whole webpage (possibly very tall) before rendering.
Why not just use Selenium?
- chromewhip uses the devtools protocol instead of the json wire protocol, where the devtools protocol has greater flexibility, especially when it comes to subscribing to granular events from the browser.
Bug reports and requests
Please simply file one using the Github tracker
Developed to run on Python 3.6, it leverages both aiohttp and asyncio for the implementation of the asynchronous HTTP server that wraps chrome.
Release history Release notifications
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size & hash SHA256 hash help||File type||Python version||Upload date|
|chromewhip-0.2.12-py3-none-any.whl (118.3 kB) Copy SHA256 hash SHA256||Wheel||3.6||Oct 18, 2018|
|chromewhip-0.2.12.tar.gz (87.9 kB) Copy SHA256 hash SHA256||Source||None||Oct 18, 2018|