This module is API-compatible with
node-phantom but doesn't rely on
WebSockets / socket.io. In essence the communication between Node and
Phantom / Slimer has been simplified significantly. It has the following advantages
over node-phantom:
Fewer dependencies/layers.
Doesn't use the unreliable and huge socket.io.
Works under cluster (node-phantom
does not, due to how it works)
server.listen(0) works in cluster.
Supports SlimerJS.
Migrating 1.x -> 2.x
Your software should work without changes, but can show deprecation warning
about outdated signatures. You need to update:
options.phantomPath -> options.path
in .create().evaluate() & .waitForSelector() -> move callback to last
position of arguments list.
That's all!
Installing
npm install node-phantom-simple# Also need phantomjs OR slimerjs:npm install phantomjs# ORnpm install slimerjs
Note. SlimerJS is not headless and requires a windowing environment.
Under Linux/FreeBSD/OSX xvfb can be used to run headlessly.. For example, if you wish
to run SlimerJS on Travis-CI, add those lines to your .travis.yml config:
You can use it exactly like node-phantom, and the entire API of PhantomJS
should work, with the exception that every method call takes a callback (always
as the last parameter), instead of returning values.
var driver = require('node-phantom-simple');driver.create({ path: require('phantomjs').path }, function (err, browser) { return browser.createPage(function (err, page) { return page.open("http://tilomitra.com/repository/screenscrape/ajax.html", function (err,status) { console.log("opened site? ", status); page.includeJs('http://ajax.googleapis.com/ajax/libs/jquery/1.7.2/jquery.min.js', function (err) { // jQuery Loaded. // Wait for a bit for AJAX content to load on the page. Here, we are waiting 5 seconds. setTimeout(function () { return page.evaluate(function
.create(options, callback)
options (not mandatory):
path (String) - path to phantomjs/slimerjs, if not set - will search in $PATH
parameters (Array) - CLI params for executed engine, [ { nave: value } ].
You can also pass in an array to use verbatim names and values.
ignoreErrorPattern (RegExp) - a regular expression that can be used to
silence spurious warnings in console, generated by Qt and PhantomJS.
On Mavericks, you can use /CoreText/ to suppress some common annoying
font-related warnings.
You can rely on globally installed engines, but we recommend to pass path explicit:
driver.create({ path: require('phantomjs').path }, callback)// or for slimerdriver.create({ path: require('slimerjs').path }, callback)
You can also have a look at the test directory to see some examples
of using the API, however the de-facto reference is the
PhantomJS documentation.
Just mentally substitute all return values for callbacks.
WebPage Callbacks
All of the WebPage callbacks have been implemented including onCallback,
and are set the same way as with the core phantomjs library:
This includes the onPageCreated callback which receives a new page object.
Properties
Properties on the WebPage
and Phantom
objects are accessed via the get()/set() method calls:
page.get('content', function (err, html) { console.log("Page HTML is: " + html);});page.set('zoomfactor', 0.25, function () { page.render('capture.png');});// You can get/set nested values easy!page.set('settings.userAgent', 'PhAnToSlImEr', callback);
Known issues
Engines are buggy. Here are some cases you should know.
.evaluate can return corrupted result:
SlimerJS: undefined -> null.
PhantomJS:
undefined -> null
null -> '' (empty string)
[ 1, undefined, 2 ] -> null
page.onConfirm() handler can not return value due async driver nature.
Use .setFn() instead: page.setFn('onConfirm', function () { return true; }).
//Get what you want from the page using jQuery. A good way is to populate an object with all the jQuery commands that you need and then return the object.