Programmatic API (experimental)

With the help of gemini/api module you can use Gemini programmatically in your scripts or build tools plugins.

First step is to create Gemini instance with a config:

var Gemini = require('gemini/api');

var gemini = new Gemini({
    projectRoot: '/path/to/project',
    gridUrl: 'http://example.com/grid',
    rootUrl: 'http://test.com'
    ...
});

• new Gemini(filePath) will load config from YAML file at given paths;

• new Gemini(options) will create config from specified options (see config file reference). Options must have projectRoot setting specified.

Accessing the config options

You can get values of options via gemini.config property:

var Gemini = require('gemini/api'),
    gemini = new Gemini('/path/to/config');

console.log(gemini.config.rootUrl);

Reading the tests

gemini.readTests(paths, options) — read all of the tests from specified paths into one suite collection.

• paths is an array of files or directories paths containing Gemini tests. If not specified, will look for tests in $projectRoot/gemini directory.

Options:

• grep — regular expression to filter suites to read. By default, all tests will be added to collection. If this option is set, only suites with name matching the pattern will be added to collection.

Returns promise which resolves to a SuiteCollection object.

Here is the example that prints all top level suite names:

var Gemini = require('gemini/api'),
    gemini = new Gemini('/path/to/config');

gemini.readTests()
    .done(function(collection) {
        collection.topLevelSuites().forEach(function(suite) {
            console.log(suite.name);
        });
    });

Suite Collection

You can create SuiteCollection object by using gemini.SuiteCollection constructor. Also SuiteCollection object is returned by gemini.readTests method.

SuiteCollection API:

• SuiteCollection([suites]) — constructor. Takes optional suites parameter, these are the top level suites.

• add(suite) — add suite to collection.

• topLevelSuites() — return array of top level suites.

• allSuites() — return array of all suites in collection. Goes through all suites children recursively.

• disableAll() — disable all suites in collection

• enableAll() — enable all suites in collection

• disable(suite, [opts]) — disable suite and all its children

  suite can be a real suite object, or suite full name

  opts are optional:

  • opts.browser — browser to disable suite in
  • opts.state — disable only specified state

• enable(suite, [opts]) — enable suite and all its children. Arguments are the same as in disable

Example on how to run only certain states in certain browsers:

var collection = gemini.readTests(paths),
    suite = findSomeSuite(collection);

collection
  .disableAll()
  .enable(suite, {state: 'some-state', browser: 'ie9'})
  .enable(suite, {state: 'other-state', browser: 'firefox'});

return gemini.test(collection);

Suite

Suite objects have the following properties:

• id — unique numeric identificator of the suite. Automatically generated when loading suites.

• name — the name of the suite.

• children — array of subsuites of the current suite.

• states — array of the State objects, defined in a suite.

State

Suite objects have the following properties:

• name — the name of the state.

Methods:

• shouldSkip(browserId) — returns true if this state should be skipped for a browser.

Updating reference screenshots

Use gemini.update(paths, options) method. By default, this command will update reference images that have diff and generate new reference images for new tests.

paths is the array of file paths or directories to run the suites from or SuiteCollection instance.

Options:

• reporters — array of reporters to use. Each element can be either string (to use corresponding built-in reporter) or reporter function (to use a custom reporter).

• grep — regular expression to filter suites to run. By default, all tests will be executed. If this option is set, only suites with name matching the pattern will be executed.

• browsers — array of browser ids to execute tests in. By default, tests are executed in all browsers, specified in config.

• sets — array of set names to execute tests in. By default, tests are executed for all sets, specified in config.

• diff(Boolean) — update only existing images with some diff, states with no reference images will be ignored.

• new(Boolean) — generate only missing images.

Returns promise that resolve to a stats object with following keys:

• total — total number of tests executed.

• skipped — number of skipped tests.

Rejects promise if critical error occurred.

Running tests

Use gemini.test(paths, options) method.

paths is the array of file paths or directories to run the tests from or SuiteCollection instance.

Options:

• reporters — array of reporter to use. Each element can be either string (to use corresponding built-in reporter) or reporter function (to use a custom reporter).

• grep — regular expression to filter suites to run. By default, all tests will be executed. If this option is set, only suites with name matching the pattern will be executed.

• browsers — array of browser ids to execute tests in. By default, tests are executed in all browsers, specified in config.

• sets — array of set names to execute tests in. By default, tests are executed for all sets, specified in config.

Returns promise that resolve to a stats object with following keys:

• total — total number of tests executed.

• skipped — number of skipped tests.

• passed — number of passed tests.

• failed — number of failed tests.

Rejects promise if critical error occurred.

Halting

gemini.halt(error, [timeout=60000ms]);

Method for abnormal termination of the test run in case of a terminal error. If process fails to gracefully shutdown in timeout milliseconds, it would be forcibly terminated (unless timeout is explicitly set to 0).

Utilites

• gemini.getScreenshotPath(suite, stateName, browserId) — returns path to the reference screenshot of the specified state for specified browser.

• gemini.getBrowserCapabilites(browserId) — returns WebDriver capabilities for specified browserId.

• gemini.browserIds — list of all browser identificators to use for tests.

• Gemini.readRawConfig — reads configuration file for specified filePath and returns content as JS object. This method does not validate and analyze gemini configuration.

Events

gemini instance emits some events, which can be used by external scripts or plugins.

See the details about available events.

Plugin example with the listening on the events:

module.exports = (gemini, options) => {
    gemini.on(gemini.events.START_RUNNER, () => {
        return setUp(gemini.config, options.param); // config can be mutated
    });

    gemini.on(gemini.events.END_RUNNER, () => {
        return tearDown();
    });
};