veltman / flipbookjs

For automatically flip-booking progress while developing something for the web.

Flipbook.js

For automatically flip-booking progress while developing something for the web.

Why?

I wanted a "set it and forget it" way to track the visual evolution of something I'm working on that I could revisit later to remember all the different design issues that came up and forks in the road.

How?

This uses PhantomJS to automatically screencap a provided URL from the command line with the option of multiple breakpoints, and saves screenshots out with timestamps and dimensions in the filename.

First, install PhantomJS.

Then, to get started, run:

phantomjs flipbook.js

This will give you a list of the extra options. The only required argument is a URL to capture, like:

phantomjs flipbook.js -u http://www.google.com

Flipbooking

You can run it at an interval with the -i option, but a cronjob is a better idea so the process isn't just running continuously. e.g.:

crontab -e
	
	Add this line: */5 * * * * phantomjs /path/to/flipbook.js -u http://localhost/site-i-am-working-on/

Now you'll start generating screenshots every 5 minutes, which can be used as slides, frames of a movie, or an animated GIF. Because everyone loves GIFs.

You'll end up with a bunch of images, named according to the dimensions of the viewport they were captured for, and the time they were taken. You can specify a folder to save them in with the -p option, or it will come up with a default folder name based on the URL. If you want to add a particular slug to the files for your reference, you can use the -f option.

There's no built-in option to stitch this stuff together, because it turns out making animated GIFs in JavaScript without a bunch of dependencies is a pain in the ass, so you're probably better off using your method of choice, or a service like: http://mothereffinganimatedgif.com/

Options

-u [url]
Url to capture. REQUIRED.
Ex: phantomjs flipbook.js -u http://www.google.com

-f [filename]
A prefix for the image filenames. By default, the format of the images created will be: [width]x[height]-[year]-[month]-[day]-[hour]-[minute].png

Ex: phantomjs flipbook.js -u http://www.google.com -f googleproject results in files like: googleproject-1024x768-2013-09-12-11-08.png

-v [viewport]
A viewport size to capture, in format "WxH". Default is 1024x768. You can list as many as you want.
Ex: phantomjs flipbook.js -u http://www.google.com -v 320x480 -v 640x480 -v 1600x1200

-c
Clip to page fold. By default, the resulting image will show the entire height of the page. With clipping on, it will clip to the window height.
Ex: phantomjs flipbook.js -u http://www.google.com -v 320x480 -c

-d [delay]
Option delay between render and capture, in milliseconds. Use this if the page has AJAX content you need to wait for. Default is 3000 (3 seconds).
Ex: phantomjs flipbook.js -u http://www.google.com -d 8000

-p [path]
Path to save images to. If not specified, it will create a new folder in the current folder with a name based on the URL.
Ex: phantomjs flipbook.js -u http://www.google.com -i folder_name
Ex: phantomjs flipbook.js -u http://www.google.com -p /path/to/save/to

-i [interval]
Run this at an interval. If it's a number it will treat it as milliseconds, but the format 5s, 5m, 5h also works (for every 5 seconds, every 5 minutes, every 5 hours, respectively). Not recommended, better to use a cronjob. Minimum value of 1 minute.
Ex: phantomjs flipbook.js -u http://www.google.com -i 5000

To Do

• Write a stitching script to scan a folder for screencaps and GIF them.
• Fix the queuing issue for the -i option. Currently it just lazily forces a reasonable minimum delay of 1 minute, because PhantomJS will only run one capture at a time, and the threading/queuing gets a little sticky. This is a lazy kludge but should work for most cases.