imclab / manet

Website screenshot service powered by Node.js, SlimerJS and PhantomJS

Manet

There is only one true thing: instantly paint what you see. When you've got it, you've got it. When you haven't, you begin again. All the rest is humbug.

Manet is a REST API server which allows capturing screenshots of websites using various parameters. It is a good way to make sure that your websites are responsive or to make thumbnails.

Manet could use different engines to work: SlimerJS or PhantomJs.

• SlimerJS runs on top of Gecko (the browser engine of Mozilla Firefox) and SpiderMonkey (the JavaScript engine of Firefox).
• PhantomJS runs on top of WebKit and JavaScriptCore.

Project was named in honor of Édouard Manet, French painter (1832-1883). He was one of the first 19th-century artists to paint modern life, and a pivotal figure in the transition from Realism to Impressionism.

Main features

• Ready-To-Use
• Supporting SlimerJS and PhantomJS
• Configurable CLI application
• Flexible REST API
• File caching
• Various image formats
• Sandbox UI

Setup

Preset

Choose and install needed engine (PhantomJS, SlimerJS, or both of them):

SlimerJS:

• You can download SlimerJS from the official site and install manually.
• or you can use the power of NPM:

npm install -g slimerjs

Gecko, the rendering engine of Firefox, cannot render web content without a graphical window, but you can launch SlimerJS with xvfb if you are under linux or MacOSx, to have a headless SlimerJS, so it is also necessary to install Xvfb (X virtual framebuffer) for *nix or OS X systems.

For example, you can use apt-get to install xvfb on Ubuntu:

sudo apt-get install xvfb

PhantomJS

• You can download PhantomJS from the official site and install manually.
• or you can also use NPM:

npm install -g phantomjs

IMPORTANT: PhantomJS is used by default (see default.json file).

Installation

After preliminaries operations you can install Manet using NPM:

npm install -g manet

That is all, now you can start and use Manet server. As you can see, it is unnecessary to clone Git repository or something else.

Server launching

Server launching is a simple as possible:

manet

If everything is OK, you should see the following message:

info: Manet server started on port 8891

Server configuration

Manet server uses hierarchical configurations to cover differnet usage use-cases:

• Command-line parameters
• Environment variables
• Built-in configuration JSON file ("config/default.json")

Rules of overriding:

• Each configuration level could be overridden by another level.
• The most-priority parameters are command-line parameters.
• The less-priority parameters are stored in build-in configuration file.

CLI parameters

--port

Web server port number. REST API and UI will be available on this port (default: 8891).

--engine

Engine for screenshot capturing: "phantomjs" or "slimerjs" (default is "slimerjs"). Specific command will be detected by configuration file (default.json) using engine parameter and OS platform.

--command

Configuration file "default.json" supports specific commands for different platforms (ex: "linux": "xvfb-run -a slimerjs"). Needed command will be detected in runtime by platform/OS. This parameter allows to override command for executing SlimerJS. It allows using full power of SlimerJS command line options to configure proxy, SSL protocol, etc. More information could be found here: http://docs.slimerjs.org/current/configuration.html
IMPORTANT: This parameter overrides "--engine" parameter.

--storage

File storage for cache (default is global temp directory).

--cache

Lifetime for file cache in seconds. Screenshots are cached for *60 minutes by default*, so that frequent requests for the same screenshot don't slow the service down. You can configure longer life for cache items or make them ethereal (use zero or negative value).

--cleanup

Remove FS storage on server startup (default is false).

--compress

Additional compression for captured screenshots using Imagemin (default is false). File sizes are significantly reduced due to this, but it requires additional processing time.

--silent

Run Manet server with or without logging information (default is false).

--level

Setting the level for your logging message. Possible values: debug, info, silly, warn, error (default is "debug").

--cors

Enable Cross-Origin Resource Sharing (default is false).

--ui

Enable or disable sandbox UI (default is true).

--timeout

Number of milliseconds to wait for the program to complete before sending it `SIGTERM` (default is 60000).

--options:{option}

Default query parameters. See also "Query parameters" for more details. Example: "--options:width 101".

Configuration file

Built-in configuration could be found in manet directory. For example, on Ubuntu it is located here: "/usr/local/lib/node_modules/manet/".

Default configuration file ("default.json"):

{
    "port": 8891,
    "cors": false,
    "ui": true,

    "silent": false,
    "level": "debug",

    "engine": "phantomjs",
    "options": {},
    "timeout": 60000,
    "compress": false,
    "cache": 3600,
    "cleanup": false,

    "commands": {
        "slimerjs": {
            "linux": "xvfb-run -a slimerjs",
            "freebsd": "xvfb-run -a slimerjs",
            "sunos": "xvfb-run -a slimerjs",
            "darwin": "slimerjs",
            "win32": "slimerjs.bat"
        },
        "phantomjs": {
            "linux": "phantomjs",
            "freebsd": "phantomjs",
            "sunos": "phantomjs",
            "darwin": "phantomjs",
            "win32": "phantomjs"
        }
    },

    "whitelist": [
        ".*"
    ]
}

CSS customization

It is possible to configure custom CSS rules, which will be applied before screenshot rendering. Use src/config/default-styles.css file to do it.

Default configuration:

body {
    background: #fff;
}

As you can see, Manet already has got pre-defined color for body background. It was made especially to prevent transparent background for some sites.

REST API

REST API is available on "/" using:

• GET method
• POST method with Content-Type:
  • application/json
  • or application/x-www-form-urlencoded

Few rules:

• The "url" parameter must be specified.
• Query parameters will be used in priority and override others.

Query parameters

url

Website address (URL). This is the only required parameter for the HTTP request. It is unnecessary for the most cases to configure scheme. Example: "github.com".

width

This property allows to change the width of the viewport, e.g., the size of the window where the webpage is displayed (default: 1024)

height

This property allows to change the height of the viewport. If width is defined and height is not defined, than full page will be captured.

clipRect

This property defines the rectangular area of the web page to be rasterized. Format: "top,left,width,height", example: "20,20,640,480".

zoom

Zoom factor of the webpage display. Setting a value to this property decreases or increases the size of the web page rendering. A value between 0 and 1 decreases the size of the page, and a value higher than 1 increases its size. 1 means no zoom (normal size). (default: 1).

quality

The compression quality. A number between 0 and 1 (default value: 1).

delay

Do a pause during the given amount of time (in milliseconds) after page opening (default: 100).

format

Indicate the file format for output image (default is "png"). Possible values: jpg, jpeg, png, gif, pdf

agent

String to define the "User-Agent" in HTTP requests. By default, it is something like:

• PhantomJS: Mozilla/5.0 (Macintosh; Intel Mac OS X) AppleWebKit/534.34 (KHTML, like Gecko) PhantomJS/1.9.0 (development) Safari/534.34
• SlimerJS: Mozilla/5.0 (X11; Linux x86_64; rv:21.0) Gecko/20100101 SlimerJS/0.7

headers

This property specifies additional HTTP request headers that will be sent to the server for every request issued (for pages and resources). Format: "key1=value1;key2=value2;..." Headers names and values get encoded in US-ASCII before being sent. Please note that setting the 'User-Agent' will overwrite the value set via "agent" parameter.

user

User name to give to HTTP Basic authentication.

password

Password to give to HTTP Basic authentication.

js

false to deactivate javascript in web pages (default is true).

images

false to deactivate the loading of images (default is true).

force

Use the force reloading of web page without using cache (default is false).

callback

Return an empty response immediately (HTTP 200 OK), then send a POST request to the callback URL when the screenshot is ready (with image in the body).

Query examples

For a quick test with the command line (using curl), type:

curl http://localhost:8891/?url=github.com > github.png
curl -H "Content-Type: application/json" -d '{"url":"github.com"}' http://localhost:8891/ > github.png
curl -H "Content-Type: application/x-www-form-urlencoded" -d 'url=github.com' http://localhost:8891/ > github.png

or (using wget)

wget http://localhost:8891/?url=github.com -O github.png

Here are some query examples that could be executed by any REST API client:

# Take a screenshot of the github.com.
GET /?url=github.com

# Custom viewport size. Return a 800x600 PNG screenshot of the github.com homepage.
GET /?url=github.com&width=800&height=600

# Clipping Rectangle. Return a screenshot clipped at [top=20, left=30, width=90, height=80]
GET /?url=github.com&clipRect=20%2C30%2C90%2C80

# Zoom rendered page in 2 times.
GET /?url=github.com&zoom=2

# Specify image output format.
GET /?url=github.com&format=jpeg

# Disable JavaScript. Return a screenshot with no JavaScript executed.
GET /?url=github.com&js=false

# Disable images. Return a screenshot without images.
GET /?url=github.com&images=false

# Custom User Agent.
GET /?url=github.com&agent=Mozilla%2F5.0+(X11%3B+Linux+x86_64)+AppleWebKit%2F537.36+(KHTML%2C+like+Gecko)+Chrome%2F34.0.1847.132+Safari%2F537.36

# HTTP Basic Authentication. Return a screenshot of a website requiring basic authentication.
GET /?url=mysite.com&user=john&password=smith

# Screenshot delay. Return a screenshot of the github.com homepage 1 second after it's loaded.
GET /?url=github.com&delay=1000

# Force page reloading. Return a screenshot without using file cache.
GET /?url=github.com&force=true

# Specify custom HTTP headers.
GET /?url=google.com&headers=User-Agent=Firefox;Accept-Charset=utf-8

# Asynchronous call.
GET /?url=github.com&callback=http://localhost:8891

Sandbox UI

Sandbox UI is available on "/" by direct GET request without "url" query parameter. It is a simple playground to build HTTP requests and try them.

Demo instance is available on Heroku: https://manet.herokuapp.com

Development

• To install project dependencies:

npm install

• To run jshint checks:

npm run lint

• To run Mocha unit tests:

# using NPM:
npm test
# using mocha and watcher:
mocha --watch -R spec

• To run Manet server:

./bin/manet

Thanks to

• SlimerJS author Laurent Jouanneau and SlimerJS community.
• PhantomJS author Ariya Hidayat and PhantomJS community.

License

The MIT License (MIT)

Copyright (c) 2014 Vladislav Bauer (see LICENSE).