aft-core

AFT-Core

the base Automated Functional Testing (AFT) library providing support for Plugins, configuration, and helper classes and functions

Installation

> npm i aft-core

Configuration

the aft-core package contains the aftConfig constant class (instance of new AftConfig()) for reading in configuration an aftconfig.json file at the project root. this configuration can be read as a top-level field using aftConfig.get('field_name') or aftConfig.get('field_name', defaultVal) and can also be set without actually modifying the values in your aftconfig.json using aftConfig.set('field_name', val). additionally, configuration classes can be read using AftConfig with the aftConfig.getSection(ConfigClass) which will read from your aftconfig.json file for a field named ConfigClass

NOTE: when a new instance of AftConfig is created the dotenv package is run and any .env file found at your project root (process.cwd()) will be processed into your environment variables making it easier to load values when developing and testing locally.

Ex: with an aftconfig.json containing:

{
    "SomeCustomClassConfig": {
        "configField1": "%your_env_var%",
        "configField2": "some-value",
        "configField3": ["foo", true, 10]
    }
}

and with the following environment variables set:

export your_env_var="an important value"

and a config class of:

export class SomeCustomClassConfig {
    configField1: string = 'default_value_here';
    configField2: string = 'another_default_value';
    configField3: Array<string | boolean | number> = ['default_val'];
    configField4: string = 'last_default_value';
}

can be accessed using an AftConfig instance as follows:

const config = aftConfig.getSection(SomeCustomClassConfig); // or new AftConfig().getSection(SomeCustomClassConfig);
config.configField1; // returns "an important value"
config.configField2; // returns "some-value"
config.configField3; // returns ["foo", true, 10] as an array
config.configField4; // returns "last_default_value"

and if you wish to entirely disregard the configuration specified in your aftconfig.json file you can use the following (still based on the above example):

const config = new AftConfig({
    SomeCustomClassConfig: {
        configField1: 'custom_value_here'
    }
});
config.configField1; // returns "custom_value_here"
config.configField2; // returns "another_default_value"
config.configField3; // returns ["default_val"] as an array
config.configField4; // returns "last_default_value"

Helpers

the aft-core package contains several helper and utility classes, interfaces and functions to make functional testing and test development easier. These include:

• rand - random string, boolean, number and uuid generation
• convert - string manipulation like Base64 encode / decode and replacement
• ellide - string elliding supporting beginning, middle and end ellipsis
• Err - a module that can run functions in a try-catch with optional logging as well as provide formatted string outputs from Error objects
• using - automatically call the dispose function of a class that implements the Disposable interface when done
• MachineInfo - get details of the host machine and user running the tests
• CacheMap - a Map implementation that stores values with expirations where expired items will not be returned and are pruned from the Map automatically. The CacheMap can also optionally store its data on the filesystem allowing for other running node processes to read from the same cache data (e.g. sharded parallel testing)
• FileSystemMap - a Map implementation that stores its values in a file on the filesystem allowing multiple node processes to share the map data or to persist the data over multiple iterations
• fileio - a constant class providing file system write and readAs<T> functions to simplify file operations
• wait - constant class providing wait.forResult<T>(...): Promise<T>, wait.forDuration(number), and wait.until(number | Date): Promise<void> functions to allow for non-thread-locking waits
• retry - constant class providing retry(retryable): Promise<T> async function that will retry a given retryable function until it succeeds or some condition such as number of attempts or elapsed time is exceeded
• verifier - see: Testing with the Verifier section below

Custom Types

aft-core also comes with some helpful types that can make building automated tests a bit easier such as:

• Action<T> - a function accepting one typed argument T and returning void
• Func<T, Tr> - a function accepting one typed argument T and returning a specified type Tr
• Class<T> - a class of type T accepting 0 or more arguments on the constructor
• ProcessingResult - a more expressive return value that can be used when you want both a boolean success and data as a result
• JsonObject - an object that can be serialised and deserialised into a Javascript Object without loss of data
• JsonKey - a value that can be used as a valid JSON object key
• JsonValue - value that can be used as a valid JSON object value
• Merge<T1, T2, T3 = {}, T4 = {}, T5 = {}, T6 = {}> - a type that can be used to create merged types (types made up of 2 or more types)

Plugins

Example Reporting Plugin

to create your own simple reporting plugin that stores all logs until the finalise function is called you would implement the code below.

NOTE: configuration for the below can be added in a object in the aftconfig.json named OnDisposeConsoleReportingPluginConfig and optionally containing values for the supported properties of the OnDisposeConsoleReportingPluginConfig class

export class OnDisposeConsoleReportingPluginConfig {
    maxLogLines: number = Infinity;
    logLevel: LogLevel = 'warn';
};
export class OnDisposeConsoleReportingPlugin extends ReportingPlugin {
    public override get logLevel(): LogLevel { return this._lvl; }
    private readonly _lvl: LogLevel;
    private readonly _logs: Map<string, Array<LogMessageData>>;
    private readonly _maxLines: number;
    constructor(aftCfg?: AftConfig) {
        super(aftCfg);
        const cfg = this.aftCfg.getSection(OnDisposeConsoleReportingPluginConfig);
        this._lvl = cfg.logLevel ?? 'warn';
        if (this.enabled) {
            this._logs = new Map<string, Array<LogMessageData>>();
        }
    }
    override initialise = async (name: string): Promise<void> => {
        if (!this._logs.has(name)) {
            this._logs.set(name, new Array<LogMessageData>());
        }
    }
    override log = async (name: string, level: LogLevel, message: string, ...data: Array<any>): Promise<void> => {
        if (this.enabled) {
            if (LogLevel.toValue(level) >= LogLevel.toValue(this.logLevel) && level != 'none') {
                const namedLogs: Array<LogMessageData> = this._logs.get(name);
                namedLogs.push({name, level, message, args: data});
                while (namedLogs.length > this.maxLogLines) {
                    namedLogs.shift();
                }
            }
        }
    }
    override submitResult = async (name: string, result: TestResult): Promise<void> => {
        /* ignore */
    }
    override finalise = async (name: string): Promise<void> => { 
        if (this.enabled) {
            const namedLogs = this._logs.get(name);
            while (namedLogs?.length > 0) {
                let data = namedLogs.shift();
                aftLogger.log({name: data.name, level: data.level, message: data.message, args: data.args});
            });
            aftLogger.log({name: this.constructor.name, level: 'debug', message: `finalised '${name}'`});
        }
    }
}

Example TestExecutionPolicyPlugin (TestRail)

export class TestRailConfig {
    username: string;
    password: string;
    url: string = 'https://you.testrail.io';
    projectId: number;
    suiteIds: Array<number> = new Array<number>();
    planId: number;
    enabled: boolean = false;
}
export class TestRailTestExecutionPolicyPlugin extends TestExecutionPolicyPlugin {
    public override get enabled(): boolean { return this._enabled; }
    private readonly _client: TestRailClient;
    private readonly _enabled: boolean;
    constructor(aftCfg?: AftConfig) {
        super(aftCfg);
        const cfg = this.aftCfg.getSection(TestRailConfig);
        this._enabled = cfg.enabled ?? false;
        if (this.enabled) {
            this._client = new TestRailClient(this.aftCfg);
        }
    }
    override shouldRun = async (testId: string): Promise<ProcessingResult> => {
        const result = await this._client.getLastTestResult(testId);
        if (result.status === 'Passed') {
            return false; // test alraedy has passing result so don't run
        }
        return true;
    }
}

Integration with javascript test frameworks

the aft-core package comes with an AftTestIntegration class which can be extended from to allow near seamless integration of AFT's reporting and test execution flow control features. AFT already has packages for integration with a few of the major test frameworks such as Jasmine, Mocha and Jest and these can be used as examples for implementing your own as needed if you are using some other test framework (NOTE: the Mocha integration also works with Cypress).

• aft-jasmine-reporter: aft-test
• aft-mocha-reporter: aft-test
• aft-jest-reporter: aft-test

Testing with the Verifier

the Verifier class and verify functions of aft-core enable testing with pre-execution filtering based on integration with external test execution policy managers via plugin packages extending the TestExecutionPolicyPlugin class (see examples above).

// jasmine spec using `aft-jasmine-reporter` package
describe('Sample Test', () => {
    it("[C1234] expect that performAction will return 'result of action'", async () => {
        const aft = new AftTest();
        const feature: FeatureObj = new FeatureObj();
        /**
         * the `verify(assertion).returns(expectation)` function
         * checks any specified `TestExecutionPolicyPlugin` implementations
         * to ensure the test should be run. It will then
         * report to any `ReportingPlugin` implementations
         * with an `TestResult` indicating the success,
         * failure or skipped status
         */
        await aft.verify(async () => await feature.performAction())
            .returns('result of action');
    });
});

in the above example, the await feature.performAction() call will only be run if a TestExecutionPolicyPlugin is loaded and returns true from it's shouldRun(testId: string) function (or no TestExecutionPolicyPlugin is loaded). additionally, any logs associated with the above verify call will use a logName of "expect_that_performAction_will_return_result_of_action" resulting in log lines like the following:

09:14:01 - [expect that performAction will return 'result of action'] - TRACE - no TestExecutionPolicyPlugin in use so run all tests

VerifierMatcher

the .returns(...) function on a Verifier can accept a VerifierMatcher instance to enhance the comparison capabilities performed by the Verifier.returns check. the following VerifierMatcher types are supported within AFT Core:

NOTE: if no VerifierMatcher is supplied then equaling is used by default

• equaling: performs a '==' test between the actual and expected. ex: await verify(() => 0).returns(equaling(false)); // success
• exactly: performs a '===' test between the actual and expected. ex: await verify(() => 0).returns(exactly(false)); // fail
• equivalent: iterates over all keys of expected and compares their type and values to those found on actual. ex: await verify(() => {foo: 'bar'}).returns({foo: 'foo'}); // fail
• between: verifies that the actual numerical value is either equal to or between the minimum and maximum expected values. ex: await verify(() => 42).returns(between(42, 45)); // success
• containing: verifies that the actual collection contains the expected value. ex: await verify(() => [0, 1, 2, 3]).returns(containing(2)); // success
• havingProps: iterates over all keys of expected and compares their type to those found on actual. this differs from equivalent in that the actual values are not part of the comparison. ex: await verify(() => {foo: 'bar'}).returns(havingProps({foo: 'foo'})); // success
• havingValue: verifies that the actual is not equal to null or undefined. ex: await verify(() => false).returns(havingValue()); // success
• greaterThan: verifies that the actual numerical value is greater than the expected. ex: await verify(() => 2).returns(greaterThan(0)); // success
• lessThan: verifies that the actual numerical value is less than the expected. ex: await verify(() => 0).returns(lessThan(1)); // success
• not: a special use VerifierMatcher that negates any VerifierMatcher passed into it. ex: await verify(() => [0, 1, 2]).returns(not(containing(1))); // fails