Contact Us 1-800-596-4880

BDD Test-Writing Syntax Reference

Behavior Driven Development (BDD) is the syntax for writing BAT tests for processing in DataWeave. BDD is an embedded domain specific language (EDSL) that resembles other testing frameworks, such as Mocha and Jasmine.

The following statement is required in each BAT test:

import * from bat::BDD

This statement imports the Domain Specific Language (DSL) testing library BDD.dwl, which defines words you frequently use when writing, such as describe, it, GET, POST, and sleep).

This statement imports common matchers, such as mustEqual, mustMatch, every, oneOf, and assert.

The following example shows how you typically begin writing a test:

import * from bat::BDD
import * from bat::Assertions
import * from bat::Mutable


var context = HashMap()
---
describe("the login flow") in [
  it("tries a secured endpoint") in [
    GET `http://127.0.0.1:4567/secured_by_token` with {

    } assert [
      $.response.status mustEqual 401,
    ]
  ],
  it("gets the access token") in [
    POST `http://127.0.0.1:4567/get_access_token` with {} assert [
      $.response.body.new_token mustMatch /accessToken_\d+/,
    ] execute [
      context.set('token', $.response.body.new_token)
    ] assert [
      context.get('token') mustMatch /accessToken_\d+/,
      context.get('token') mustEqual $.response.body.new_token,
      $.response.status mustEqual 200,
    ]
  ],
  it("tries to get a secured endpoint") in [
    GET `http://127.0.0.1:4567/secured_by_token/header` with {
      headers: {
        Authorization: context.get('token')
      }
    } assert [
      $.request.headers.Authorization mustEqual context.get('token'),
      $.response.status mustEqual 200
    ]
  ]
]

About Test Files

Create a .dwl file for each functional scenario. A folder-based organization of files is recommended. This organization accommodates reports the BAT testing generates. For example:

tests
├── CoreServices
│   └── Sanity.dwl
└── DesignCenter
    ├── APIDesigner
    │   └── Designer.dwl
    └── MyFolder
        ├── DataSense.dwl
        └── GetProjects.dwl

A test file must contain a describe expression.

import * from bat::BDD
---
describe `Demo scenario` in [

]

A test suite must contain at least one test (.dwl) file.

Keeping Confidential Information Secure in Tests

In your tests, you can access shared secrets from Anypoint Secrets Manager. Secrets Manager enables you to securely store sensitive information, such as a password, authentication token, endpoint URL, or webhook URL.

Limitations

  • Functional Monitoring works only with symmetric key shared secret types from Anypoint Secrets Monitor.

  • You can use aliases for shared secrets in a test suite only if you plan to run that test suite from a private location.

Configure a Secret in Anypoint Secret Manager

To keep sensitive information secret in tests you run from a private location, create a shared secret in Secrets Manager.

To configure a shared secret:

  1. Open a secrets group or create a new one. Create the group in the same environment that you are using in BAT CLI.

    To find out which environment you are using in BAT CLI:

    1. Run the command bat whoami.

      The output has the ID for the environment.

    2. Run the command bat environment ls to list the environments to which you have access.

    3. Match the ID from the bat whoami command with one of the environments listed.

    4. If you need to switch to the environment that your secrets group is in, run the command bat environment switch name, where name is the name of the environment.

  2. Select Shared Secret.

  3. In the Type field, select Symmetric Key.

  4. In the Key field, paste the sensitive information encoded as a Base64 string.

  5. Paste the Base64 string into the Confirm Key field.

After you modify the monitor’s main.dwl file to add the secret information to the code, grant the monitor access to the secrets you configured in Secrets Manager. See Grant Your Monitor Access to the Secret.

Grant Your Monitor Access to the Secret

After you modify your monitor to add the secret information, grant the monitor access to the secret. At runtime, BAT CLI looks up the shared secrets using the aliases.

To grant a monitor access to the secret:

  1. Copy the name of the new shared secret obtained in Configure a Secret in Anypoint Secret Manager.

  2. At a command prompt, run the bat grant command, specifying an alias for the shared secret. When you run this command, BAT CLI creates a section named secrets in your test suite’s bat.yaml file, if the section does not already exist. In that section, BAT CLI adds these indented lines:

    alias:
     secretId: "secret-ID"
    • alias: The alias that you specified in the bat grant command.

    • secret-ID: The ID of the secret within Anypoint Secrets Manager. This ID does not appear in ASM, so there is no way for someone looking in your bat.yaml file to associate the ID with any particular secret. The BAT CLI uses this ID to look up the secret that you associated with the alias.

Example of Using an Alias

Use the alias in a test by storing it in a variable, as in this example:

import * from bat::BDD
import * from bat::Assertions

var mySecret: String = secret('myAlias') default 'notFound'

---
suite(parts="Simple Suite") in [
  it must 'answer 200' in [
    GET '$(config.url)?token=$(mySecret)' with {} assert [
      $.response.status mustEqual 200
    ] execute
  ]
]
mySecret

The variable in which the secret is stored in this example. The variable is declared as a String.

secret('myAlias')

The secret() function requires the input 'name-of-alias', where name-of-alias is the alias of the shared secret that you want to use.

default 'notFound'

The message to show if a shared secret with the specified alias does not exist in Anypoint Secrets Manager.

$(mySecret)

The use of the variable mySecret to supply the authentication token for the endpoint at the URL that is defined in the configuration file that was specified when the bat command to run the test suite was entered.

For example, suppose that the configuration file was this:

The example configuration file dev-environment.dwl
{
  url: "https://dev.yourcompany.com/shoppingCart"
}

In this case, the command to run this test suite, when the command is run in the directory in which the bat.yaml file for the test suite is located and the default profile is used, is this:

bat --config=dev-environment.dwl

At runtime, the BAT CLI replaces the variable with the shared secret that it finds associated with the specified alias in Anypoint Secrets Manager.

Test Blocks

Test blocks make the test more readable. You can write every test block as a function call or a custom interpolation.

Test Keywords

You use the following keywords and follow Fowler’s GivenWhenThen structure, in tests:

Function Name Description Stops execution if it fails

describe

For grouping, which you can nest as deep to add contextual information for reports

No

scenario

Same as describe

No

suite

Same as describe

No

given

Test step, used for preparation

Yes

when

Test step, normally used to create side effects

Yes

it

Test step, validations

No

should

Test step

No

must

Test step

Yes

it should "str"

Test step

No

it must ""

Test step

Yes

while

Test step

No

until

Test step

No

About a Test Block

Test blocks have a shorthand name to skip the block described as x<name>. For example, the skipped version of describe is xdescribe.

describe("A flow")  in [ ... ] // Describes a sequence of test blocks
xdescribe("A flow") in [ ... ] // The same, but skipped.

describe `A flow` in [ ... ]   // The same as describe, but with a custom interpolation
xdescribe `A flow` in [ ... ]  // The same, but skipped.

About Custom Interpolators

Custom interpolators provide a way to call functions and decompose interpolations in two lists:

  • The first list contains all the string literals.

  • The second one contains the interpolated values.

You do not need to convert the values to strings. Enclose a custom interpolation text and value in backticks (`).

Every item in the following block can be used as custom interpolators and has the x<name> version.

import * from bat::BDD         // <-----
import * from bat::Assertions
---
describe `User trades stocks` in [
  scenario `User requests a sell before close of trading` in [
    given `I have 100 shares of MSFT stock` in [
      POST `http://broker/create_stocks` with {
        body: {
          quantity: 100,
          paper: 'MSFT'
        }
      } assert [
        $.response.status == 201
      ]
    ],
    given `I have 150 shares of APPL stock` in [
      POST `http://broker/create_stocks` with {
        body: {
          quantity: 150,
          paper: 'APPL'
        }
      } assert [
        $.response.status == 201
      ]
    ],
    when `I ask to sell 20 shares of MSFT stock` in [
      POST `http://broker/sell_stocks` with {
        body: {
          quantity: 20,
          paper: 'MSFT'
        }
      } assert [
        $.response.status == 201
      ]
    ],
    it should "have 80 shares of MSFT stock" in [
      GET `http://broker/get_stocks/MSFT` with {
        headers: {}
      } assert [
        $.response.status == 200,
        $.response.body.quantity == 80
      ]
    ],
    it should "have 150 shares of APPL stock" in [
      GET `http://broker/get_stocks/APPL` with {
        headers: {}
      } assert [
        $.response.status == 200,
        $.response.body.quantity == 150
      ]
    ]
  ]
]

You can also write the block without custom interpolators to conform to your coding style guide. This doesn’t affect behavior:

import * from bat::BDD         // <-----
import * from bat::Assertions
---
describe("User trades stocks") in [
  scenario("User requests a sell before close of trading") in [
    given("I have 100 shares of MSFT stock") in [
      POST `http://broker/create_stocks` with {
        body: {
          quantity: 100,
          paper: 'MSFT'
        }
      } assert [
        $.response.status == 201
      ]
    ],
    given("I have 150 shares of APPL stock") in [
      POST `http://broker/create_stocks` with {
        body: {
          quantity: 150,
          paper: 'APPL'
        }
      } assert [
        $.response.status == 201
      ]
    ],
    when("I ask to sell 20 shares of MSFT stock") in [
      POST `http://broker/sell_stocks` with {
        body: {
          quantity: 20,
          paper: 'MSFT'
        }
      } assert [
        $.response.status == 201
      ]
    ],
    should("have 80 shares of MSFT stock") in [
      GET `http://broker/get_stocks/MSFT` with {
        headers: {}
      } assert [
        $.response.status == 200,
        $.response.body.quantity == 80
      ]
    ],
    should("have 150 shares of APPL stock") in [
      GET `http://broker/get_stocks/APPL` with {
        headers: {}
      } assert [
        $.response.status == 200,
        $.response.body.quantity == 150
      ]
    ]
  ]
]

Controlling Execution and Deleting Assets

If you get a failure in the middle of a test, you need to delete any created assets. For example, the user creates an asset, performs a validation, and then deletes it. Typically, that validation fails, and since it breaks the test, the asset is not deleted and your database starts accumulating test data. Typically this means changing a must to a should so that execution can continue, but not leave a failed asset.

Incorrect Example:

If the update name in line 6 fails, the project is not deleted because the execution stops.

describe `update project names` in [
  it must 'create a project' in [
    createProject()     // OK
  ],
  it must 'update the name' in [
    updateProjectName() // FAILS
  ],
  // Because the previous step is a `must` that failed, execution stops here and the next steps don't execute
  it must 'clean up deleting the project' in [
    deleteProject()     // CANCELLED
  ]
]

Use these reserved words to wrap steps to stop or allow execution to continue:

  • should means something might fail, but it is not mandatory for the test.

  • must means that something, such as project creation, must execute to continue.

Correct Example:

describe `update project names` in [
  it must 'create a project' in [
    /**
     * Project creation is a MUST, because in this scenario
     * we depend on the created project to continue.
     */
    createProject()     // OK
  ],
  it should 'update the name' in [
    /**
     * Validations are should because the execution must continue
     * if the validation fails.
     */
    updateProjectName() // FAILS
  ],
  // Because the previous step is a should and it failed, continue executing.
  it must 'clean up deleting the project' in [
    deleteProject()     // OK
  ]
]

Executing Steps Selectively

The assuming function skips the test if the result is false. This command has the following syntax:

[TestBlockExpression] assuming [BooleanExpression] in …​

For example:

describe `E2E Scenario` in [
  it should 'always do something' in [
    doSomething()
  ],
  it should 'do something else' in [
    doSomethingElse()
  ],
  it should 'sometimes, do something else' assuming (random() > 0.5) in [
    // This is executed randomly, based on          ^^^^^^^^^^^^^^^^ that condition
    doSomethingElse()
  ],
  it should 'do something in dev environments' assuming (config.env == 'DEV') in [
    // This is executed only when                       ^^^^^^^^^^^^^^^^^^^^^ that == true
    doSomethingElse()
  ]
]

To make the code more readable and understandable, you can use two aliases for this function. when and whenNot. For example:

describe `E2E Scenario` in [
  it should 'always do something' in [
    doSomething()
  ],
  it must 'do something else' when config.runSanity in [
    doSomethingElse()
  ],
  it should 'do something else' when a == b in [
    doSomethingElse()
  ],
  it should 'do something in dev environments' whenNot config.isSmokeTests in [
    doSomethingElse()
  ]
]

Executing Loop Sentences

The while or until functions run the test while (or until) a condition becomes true or false.

Signature of the while function:

  • while( sentence , condition, time per request, number of retries)

  • do { sentence } while (condition) //→ Default values: 1 second and 3 retries.

Signature for the until function:

  • until( sentence , condition, time per request, number of retries)

  • do { sentence } until (condition) //→ Default values: 1 second and 3 retries.

Example:

dwl
import * from bat::BDD
import * from bat::Assertions
import * from bat::Types
---
suite("Example for until and while") in [
  it should 'test the while prefix' in [
    while(() -> GET `http://apimon.cloudhub.io/users` with {},
    (x: BATHttpStep) -> x.result.response.status != 200, 10000, 5)
  ],
  it should 'test the while infix' in [
    do {
      () -> GET `http://apimon.cloudhub.io/users` with {} assert [
          $.response.status mustEqual 200
        ]
    } while ($.response.status != 200)
  ],
  it should 'test the until prefix' in [
    until(() -> GET `http://apimon.cloudhub.io/users` with {},
     (x: BATHttpStep) -> x.result.response.status != 200, 10000, 5)
  ],
  it should 'test the until infix' in [
    do {
      () -> GET `http://apimon.cloudhub.io/users` with {} assert [
          $.response.status mustEqual 200
        ]
    } until ($.result.response.status != 200)
  ]
]