phantom-snap
Render HTML to an image using PhantomJS with this library designed to scale for high volume continuous operation.
Features
- Provides full timing control around the rendering process.
- Maintains a live PhantomJS process (instead of a new one per request which many wrappers do, which is slow).
- Render content from a URL, or provide the HTML content directly to the renderer
Roadmap
- Decorator to manage rendering under specified timezones
- Decorator to manage rendering under specified proxies.
Examples
The example assumes you have http://phantomjs.org/ installed.
This first example demonstrates rendering a URL and saving the resulting image to a file at /tmp/google-render.jpg.
from phantom_snap.settings import PHANTOMJS
from phantom_snap.phantom import PhantomJSRenderer
from phantom_snap.imagetools import save_image
config = {
'executable': '/usr/local/bin/phantomjs',
'args': PHANTOMJS['args'] + ['--disk-cache=false', '--load-images=true'],
'env': {'TZ': 'America/Los_Angeles'}
'timeouts': {
'page_load': 3
}
}
r = PhantomJSRenderer(config)
url = 'http://www.google.com'
try:
page = r.render(url, img_format='JPEG')
save_image('/tmp/google-render', page)
finally:
r.shutdown(15)
A sample response from r.render(url) looks like this:
{
"status": "success",
"format": "PNG",
"url": "http://www.google.com",
"paint_time": 141,
"base64": "iVBORw0KGgo <SNIP> RK5CYII=",
"error": null,
"load_time": 342
}
This example shows how to provide HTML content directly to the rendering process, instead of requesting it.
from phantom_snap.settings import PHANTOMJS
from phantom_snap.phantom import PhantomJSRenderer
from phantom_snap.imagetools import save_image
config = {
'executable': '/usr/local/bin/phantomjs',
'args': PHANTOMJS['args'] + ['--disk-cache=false', '--load-images=true']
}
r = PhantomJSRenderer(config)
url = 'http://www.a-url.com'
html = '<html><body>Boo ya!</body></html>'
try:
page = r.render(url=url, html=html, img_format='PNG')
save_image('/tmp/html-render', page)
finally:
r.shutdown(15)
If you would like to offload the running of phantomjs into AWS Lambda, you can use the LambdaRenderer class in the following way:
from phantom_snap.lambda_renderer import LambdaRenderer
from phantom_snap.imagetools import save_image
config = {
'url': 'http://url-to-my-lambda-func',
}
r = LambdaRenderer(config)
url = 'http://www.youtube.com'
page = r.render(url, img_format='JPEG')
save_image('/tmp/youtube-render', page)
r.shutdown()
To learn more about offloading renders into AWS Lambda, please see the serverless folder.
Decorators
Lifetime
If you plan on running a PhantomJSRenderer instance for an extended period of time with high volume, it's recommended that you wrap the instance with a Lifetime decorator as shown below.
The Lifetime decorator will transparently shutdown the underlying PhantomJS process if the renderer is idle or after a maximum lifetime to release any accumulated resources. This is important if PhantomJS is configured to use a memory-based browser cache to prevent the cache from growing too large. After the Lifetime decorator shuts down the Renderer (due to idle time or maximum time) the next render request will automatically create a new PhantomJS process.
from phantom_snap.settings import PHANTOMJS
from phantom_snap.phantom import PhantomJSRenderer
from phantom_snap.decorators import Lifetime
config = {
'executable': '/usr/local/bin/phantomjs',
'args': PHANTOMJS['args'] + ['--disk-cache=false', '--load-images=true'],
'env': {'TZ': 'America/Los_Angeles'},
# Properties for the Lifetime decorator
'idle_shutdown_sec': 900, # 15 minutes, Shutdown PhantomJS if it's been idle this long
'max_lifetime_sec': 43200 # 12 hours, Restart PhantomJS every 12 hours
}
r = Lifetime(PhantomJSRenderer(config))
try:
urls = [] # Some endless source of URL targets
for url in urls:
page = r.render(url=url, img_format='JPEG')
# Store the image somewhere
finally:
r.shutdown()
You can view the default configuration values in phantom_snap.settings.py.
