selenium-html-js-converter-lb
v0.4.4
Published
Convert selenium html to javascript (selenium-webdriver) for loadbooster
Downloads
4
Maintainers
Readme
selenium-html-js-converter
A small tools to convert 'selenium-IDE' HTML test cases into javascript test cases with wd-sync.
Install
$ npm install selenium-html-js-converter
Usage
var converter = require("selenium-html-js-converter");
converter.convertHtmlFileToJsFile(YourHtmlFile, OutJsFile[, options]);
Before you run the javascript, you need to install wd-sync, which is used in the test case.
$ npm install wd-sync
To run the javascript
var wdSync = require('wd-sync');
var client = wdSync.remote('127.0.0.1', 8910); // phantomjs default wd port
var browser = client.browser;
var sync = client.sync;
var test1 = require('./OutJsFile');
sync(function(){
browser.init({ browserName: 'phantomjs' });
test1(browser);
browser.quit();
});
Refer to wd-sync API.
If you want to run the javascript example above successfully, you should start a phantomjs webdriver server in another process.
$ phantomjs --wd
API Usage
convertHtmlFileToJsFile(filePath, filePath[, options]) => void;
convertHtmlStrToJsFile(htmlStr, filePath[, options]) => void;
convertHtmlFileToJsStr(filePath[, options]) => string;
convertHtmlStrToJsStr(htmlStr[, options]) => string;
convertHtmlSuiteFileToJsFiles(htmlFile[, options]) => void;
Options
The options argument is optional, and for backwards compatibility, it may be a string. When it is a string, it sets options.outputFolder
for convertHtmlSuiteFileToJsFiles()
and options.testCaseName
for all other methods.
The argument is an object hash and may contain the following properties:
General options
timeout
Sets the default timeout for waits and retries, in msecs. The default is 30,000 msecs.
retries
Sets the number of retries that should be attempted when a test fails. This may help prevent false negatives.
When set, the wait between retries is calculated to fit within the timeout
period, while increasing the wait period between retries after each failed attempt. The calculation is simple: Before the last allowed retry, there will be a wait period of timeout/2
msecs. Before the second-to-last allowed retry, there's a wait of half of that. This halving of continues up until the first retry. The wait after the very first try is the same as the wait after the first retry. In other words, given a timeout of 1,000 msecs and a maximum of five retries, you will get pauses of [ 32, 32, 63, 125, 250, 500 ]
msecs after each failed attempt, after which the test ultimately fails.
Note that the way retrying is implemented is to wrap each test in an outer function, so it will result in slightly uglier code, e.g. this test without retrying:
assert.strictEqual(!!browser.hasElementByCssSelector("[name=some-name]"), true, 'Assertion error: Expected: true, got: ' + browser.hasElementByCssSelector("[name=some-name]") + " [ Command: assertElementPresent | css=[name=some-name] ]");
becomes, when enabling retries:
retry(browser, function () {
assert.strictEqual(!!browser.hasElementByCssSelector("[name=some-name]"), true, 'Assertion error: Expected: true, got: ' + browser.hasElementByCssSelector("[name=some-name]") + " [ Command: assertElementPresent | css=[name=some-name] ]");
}, options.retries, options.timeout);
baseUrl
The base url is by default extracted from <link rel="selenium.base" href="${baseURL}" />
from the header of the test files. However, you may want to ignore this value and specify it yourself at conversion time. The Firefox Selenium IDE sets this property when it creates a new text file, and you cannot change it in the IDE afterwards. This is unfortunate if, for instance, you collect test cases from several colleagues, and their selenium.base
values are either wrong or differing between them.
JS Beautify
The generated code is formatted with JS Beautify. You may override the default formatting options by passing your own via this member object. See the JS Beautify options section for syntax and available options.
convertHtml(File|Str)ToJs(File|Str) options
testCaseName
If omitted, the test case name defaults to "UntitledTest" when converting string-to-string. When converting to files, the default is to construct the name based on the target file name.
If provided, the test case name is used to place any screenshots that may be taken during a test case. Screenshots are saved as ./screenshots/<testCaseName>/<fileName as specified in the HTML test case or an incrementing number>.png
.
Paths from the screenshot file name in the HTML test case itself are stripped, but you can include paths in the testCaseName if you want your screenshots ordered in folders, e.g.
convertHtmlFileToJsFile('test.html', 'test.js', 'backend/user management/adding and deleting');
This example will put screenshots in ./screenshots/backend/user management/adding and deleting/
. Any folders missing in the path will be automatically created when you run your test (see Run-time options on how to override this when running the tests).
Note that kwArgs aren't supported by wd in screenshots (see Selenium reference docs) so they will be ignored.
convertHtmlSuiteFileToJsFiles options
outputFolder
Suite file conversion puts each converted file in the specified output folder. Any missing folders in the output path will be created. For test cases that exist in subfolders relative to the suite file, the subfolders are appended to the corresponding js file's target path. If you do not specify an output folder, js files will be created next to each of the source html test case files.
Run-time options
You may redefine some defaults and override some test settings at run-time by defining them via properties in the second, and optional, options argument hash when executing a test case.
screenshotFolder
The screenshot folder can be changed at runtime by passing it to the test function, e.g. using the example from before:
sync(function(){
browser.init({ browserName: 'phantomjs' });
test1(browser, {
screenshotFolder: 'screenshots/firefox/backend/user management/adding and deleting'
});
browser.quit();
});
timeout
The default timeout for wait operations, e.g. waitForElementPresent, is 30 seconds. You may change this by passing the desired timeout in miliseconds as an integer, e.g.
sync(function(){
browser.init({ browserName: 'phantomjs' });
test1(browser, {
screenshotFolder: 'screenshots/firefox/backend/user management/adding and deleting',
timeout: 5000
});
browser.quit();
});
baseUrl
Changes the baseUrl for relative links. This is useful, for instance, if you test locally against a localhost version of a website, while testing with Sauce Labs against a staging server.
forceBaseUrl
Forces urls to be prefixed by the baseUrl, even when they aren't relative. Useful if you have test cases with absolute URLs and you need to force them to target a different domain. The domain part of the url will be replaced with the value of baseUrl.
Please note that (currently) testing against a url with an asterisk or question mark may cause domain substition not to work with some commands, such as assertLocation.
Logging
selenium-html-js-converter is by default quite verbose. You may redirect logging to your own logger object by calling .setLogger()
once imported, e.g.:
var converter = require("selenium-html-js-converter");
converter.setLogger({
log : function () {},
debug : function () {},
info : function () {},
warn : console.warn,
error : console.error
});
Custom commands
Bundled with the package is, so far, one custom Selenium command. In the extensions
folder is the user extensions to be loaded into the Selenium IDE (go to Settings > General > Selenium Core extensions
and browse for the file).
setWindowSize
This command will be converted and applied directly on the webdriver browser object, so it'll work regardless of browser profile and security settings. It probably won't have any effect in the Firefox Selenium IDE, though, as security settings prevents the resizing of any window via Javascript that wasn't also opened via Javascript. So, only popups will be able to be resized in the IDE, and using popups is also discouraged due to the same security settings preventing them in most other browsers (i.e. your tests would then most likely fail later when run as Javascript).
Pass the dimensions as the first argument, as width times height. Width and height can be separated by anything non-numerical, so setWindowSize('450 800')
, setWindowSize('450, 800')
, setWindowSize('450x800')
, and setWindowSize('450 pixels on the wide side and 800 pixels on the tall one, please')
are all valid arguments.