Pact Consumer is used by projects that are consumers of an API.
Most projects will want to use pact-consumer via one of the test framework specific projects. If your favourite framework is not implemented, this module should give you all the hooks you need.
Provides a DSL for use with Java to build consumer pacts.
The library is available on maven central using:
- group-id =
au.com.dius.pact - artifact-id =
consumer - version-id =
4.2.x
Example in a JUnit test:
import au.com.dius.pact.model.MockProviderConfig;
import au.com.dius.pact.model.RequestResponsePact;
import org.apache.http.entity.ContentType;
import org.jetbrains.annotations.NotNull;
import org.junit.Test;
import java.io.IOException;
import java.util.HashMap;
import java.util.Map;
import static au.com.dius.pact.consumer.ConsumerPactRunnerKt.runConsumerTest;
import static org.junit.Assert.assertEquals;
public class PactTest {
@Test
public void testPact() {
RequestResponsePact pact = ConsumerPactBuilder
.consumer("Some Consumer")
.hasPactWith("Some Provider")
.uponReceiving("a request to say Hello")
.path("/hello")
.method("POST")
.body("{\"name\": \"harry\"}")
.willRespondWith()
.status(200)
.body("{\"hello\": \"harry\"}")
.toPact();
MockProviderConfig config = MockProviderConfig.createDefault();
PactVerificationResult result = runConsumerTest(pact, config, new PactTestRun() {
@Override
public void run(@NotNull MockServer mockServer) throws IOException {
Map expectedResponse = new HashMap();
expectedResponse.put("hello", "harry");
assertEquals(expectedResponse, new ConsumerClient(mockServer.getUrl()).post("/hello",
"{\"name\": \"harry\"}", ContentType.APPLICATION_JSON));
}
});
if (result instanceof PactVerificationResult.Error) {
throw new RuntimeException(((PactVerificationResult.Error)result).getError());
}
assertEquals(PactVerificationResult.Ok.INSTANCE, result);
}
}The DSL has the following pattern:
.consumer("Some Consumer")
.hasPactWith("Some Provider")
.given("a certain state on the provider")
.uponReceiving("a request for something")
.path("/hello")
.method("POST")
.body("{\"name\": \"harry\"}")
.willRespondWith()
.status(200)
.body("{\"hello\": \"harry\"}")
.uponReceiving("another request for something")
.path("/hello")
.method("POST")
.body("{\"name\": \"harry\"}")
.willRespondWith()
.status(200)
.body("{\"hello\": \"harry\"}")
.
.
.
.toPact()You can define as many interactions as required. Each interaction starts with uponReceiving followed by willRespondWith.
The test state setup with given is a mechanism to describe what the state of the provider should be in before the provider
is verified. It is only recorded in the consumer tests and used by the provider verification tasks.
The body method of the ConsumerPactBuilder can accept a PactDslJsonBody, which can construct a JSON body as well as define regex and type matchers.
For example:
PactDslJsonBody body = new PactDslJsonBody()
.stringType("name")
.booleanType("happy")
.hexValue("hexCode")
.id()
.ipAddress("localAddress")
.numberValue("age", 100)
.timestamp();The following matching methods are provided with the DSL. In most cases, they take an optional value parameter which will be used to generate example values (i.e. when returning a mock response). If no example value is given, a random one will be generated.
| method | description |
|---|---|
| string, stringValue | Match a string value (using string equality) |
| number, numberValue | Match a number value (using Number.equals)* |
| booleanValue | Match a boolean value (using equality) |
| stringType | Will match all Strings |
| numberType | Will match all numbers* |
| integerType | Will match all numbers that are integers (both ints and longs)* |
| decimalType | Will match all real numbers (floating point and decimal)* |
| booleanType | Will match all boolean values (true and false) |
| stringMatcher | Will match strings using the provided regular expression |
| timestamp | Will match string containing timestamps. If a timestamp format is not given, will match an ISO timestamp format |
| date | Will match string containing dates. If a date format is not given, will match an ISO date format |
| time | Will match string containing times. If a time format is not given, will match an ISO time format |
| ipAddress | Will match string containing IP4 formatted address. |
| id | Will match all numbers by type |
| hexValue | Will match all hexadecimal encoded strings |
| uuid | Will match strings containing UUIDs |
| includesStr | Will match strings containing the provided string |
| equalsTo | Will match using equals |
| matchUrl | Defines a matcher for URLs, given the base URL path and a sequence of path fragments. The path fragments could be strings or regular expression matchers |
| nullValue | Matches the JSON Null value |
* Note: JSON only supports double precision floating point values. Depending on the language implementation, they may be parsed as integer, floating point or decimal numbers.
Lots of the time you might not know the number of items that will be in a list, but you want to ensure that the list
has a minimum or maximum size and that each item in the list matches a given example. You can do this with the arrayLike,
minArrayLike and maxArrayLike functions.
| function | description |
|---|---|
eachLike |
Ensure that each item in the list matches the provided example |
maxArrayLike |
Ensure that each item in the list matches the provided example and the list is no bigger than the provided max |
minArrayLike |
Ensure that each item in the list matches the provided example and the list is no smaller than the provided min |
For example:
DslPart body = new PactDslJsonBody()
.minArrayLike("users")
.id()
.stringType("name")
.closeObject()
.closeArray();This will ensure that the users list is never empty and that each user has an identifier that is a number and a name that is a string.
If the order of the list items is not known, you can use the unorderedArray matcher functions. These will match the
actual list against the expected one, except will match the items in any order.
| function | description |
|---|---|
unorderedArray |
Ensure that the list matches the provided example, ignoring the order |
unorderedMinArray |
Ensure that the list matches the provided example and the list is not smaller than the provided min |
unorderedMaxArray |
Ensure that the list matches the provided example and the list is no bigger than the provided max |
unorderedMinMaxArray |
Ensure that the list matches the provided example and the list is constrained to the provided min and max |
The array contains matcher functions allow you to match the actual list against a list of required variants. These work by matching each item against the variants, and the matching succeeds if each variant matches at least one item. Order of items in the list is not important.
The variants can have a totally different structure, and can have their own matching rules to apply. For an example of how these can be used to match a hypermedia format like Siren, see Example Pact + Siren project.
| function | description |
|---|---|
arrayContaining |
Matches the items in an array against a number of variants. Matching is successful if each variant occurs once in the array. Variants may be objects containing matching rules. |
.arrayContaining("actions")
.object()
.stringValue("name", "update")
.stringValue("method", "PUT")
.matchUrl("href", "http://localhost:9000", "orders", regex("\\d+", "1234"))
.closeObject()
.object()
.stringValue("name", "delete")
.stringValue("method", "DELETE")
.matchUrl("href", "http://localhost:9000", "orders", regex("\\d+", "1234"))
.closeObject()
.closeArray()For cases where you are expecting basic JSON values (strings, numbers, booleans and null) at the root level of the body
and need to use matchers, you can use the PactDslJsonRootValue class. It has all the DSL matching methods for basic
values that you can use.
For example:
.consumer("Some Consumer")
.hasPactWith("Some Provider")
.uponReceiving("a request for a basic JSON value")
.path("/hello")
.willRespondWith()
.status(200)
.body(PactDslJsonRootValue.integerType())If the root of the body is an array, you can create PactDslJsonArray classes with the following methods:
| function | description |
|---|---|
arrayEachLike |
Ensure that each item in the list matches the provided example |
arrayMaxLike |
Ensure that each item in the list matches the provided example and the list is no bigger than the provided max |
arrayMinLike |
Ensure that each item in the list matches the provided example and the list is no smaller than the provided min |
For example:
PactDslJsonArray.arrayEachLike()
.date("clearedDate", "mm/dd/yyyy", date)
.stringType("status", "STATUS")
.decimalType("amount", 100.0)
.closeObject()This will then match a body like:
[ {
"clearedDate" : "07/22/2015",
"status" : "C",
"amount" : 15.0
}, {
"clearedDate" : "07/22/2015",
"status" : "C",
"amount" : 15.0
}, {
"clearedDate" : "07/22/2015",
"status" : "C",
"amount" : 15.0
} ]For the case where you have arrays of arrays (GeoJSON is an example), the following methods have been provided:
| function | description |
|---|---|
eachArrayLike |
Ensure that each item in the array is an array that matches the provided example |
eachArrayWithMaxLike |
Ensure that each item in the array is an array that matches the provided example and the array is no bigger than the provided max |
eachArrayWithMinLike |
Ensure that each item in the array is an array that matches the provided example and the array is no smaller than the provided min |
For example (with GeoJSON structure):
new PactDslJsonBody()
.stringType("type","FeatureCollection")
.eachLike("features")
.stringType("type","Feature")
.object("geometry")
.stringType("type","Point")
.eachArrayLike("coordinates") // coordinates is an array of arrays
.decimalType(-7.55717)
.decimalType(49.766896)
.closeArray()
.closeArray()
.closeObject()
.object("properties")
.stringType("prop0","value0")
.closeObject()
.closeObject()
.closeArray()This generated the following JSON:
{
"features": [
{
"geometry": {
"coordinates": [[-7.55717, 49.766896]],
"type": "Point"
},
"type": "Feature",
"properties": { "prop0": "value0" }
}
],
"type": "FeatureCollection"
}and will be able to match all coordinates regardless of the number of coordinates.
The DSL has been extended for cases where the keys in a map are IDs. For an example of this, see
#313. In this case you can use the eachKeyLike method, which takes an
example key as a parameter.
For example:
DslPart body = new PactDslJsonBody()
.object("one")
.eachKeyLike("001", PactDslJsonRootValue.id(12345L)) // key like an id mapped to a matcher
.closeObject()
.object("two")
.eachKeyLike("001-A") // key like an id where the value is matched by the following example
.stringType("description", "Some Description")
.closeObject()
.closeObject()
.object("three")
.eachKeyMappedToAnArrayLike("001") // key like an id mapped to an array where each item is matched by the following example
.id("someId", 23456L)
.closeObject()
.closeArray()
.closeObject();For an example, have a look at WildcardKeysTest.
You can use regular expressions to match incoming requests. The DSL has a matchPath method for this. You can provide
a real path as a second value to use when generating requests, and if you leave it out it will generate a random one
from the regular expression.
For example:
.given("test state")
.uponReceiving("a test interaction")
.matchPath("/transaction/[0-9]+") // or .matchPath("/transaction/[0-9]+", "/transaction/1234567890")
.method("POST")
.body("{\"name\": \"harry\"}")
.willRespondWith()
.status(200)
.body("{\"hello\": \"harry\"}")You can use regular expressions to match request and response headers. The DSL has a matchHeader method for this. You can provide
an example header value to use when generating requests and responses, and if you leave it out it will generate a random one
from the regular expression.
For example:
.given("test state")
.uponReceiving("a test interaction")
.path("/hello")
.method("POST")
.matchHeader("testreqheader", "test.*value")
.body("{\"name\": \"harry\"}")
.willRespondWith()
.status(200)
.body("{\"hello\": \"harry\"}")
.matchHeader("Location", ".*/hello/[0-9]+", "/hello/1234")You can use regular expressions to match request query parameters. The DSL has a matchQuery method for this. You can provide
an example value to use when generating requests, and if you leave it out it will generate a random one
from the regular expression.
For example:
.given("test state")
.uponReceiving("a test interaction")
.path("/hello")
.method("POST")
.matchQuery("a", "\\d+", "100")
.matchQuery("b", "[A-Z]", "X")
.body("{\"name\": \"harry\"}")
.willRespondWith()
.status(200)
.body("{\"hello\": \"harry\"}")By default, when the pact file is written, it will be merged with any existing pact file. To force the file to be
overwritten, set the Java system property pact.writer.overwrite to true.
You can have values from the provider state callbacks be injected into most places (paths, query parameters, headers, bodies, etc.). This works by using the V3 spec generators with provider state callbacks that return values. One example of where this would be useful is API calls that require an ID which would be auto-generated by the database on the provider side, so there is no way to know what the ID would be beforehand.
The following DSL methods allow you to set an expression that will be parsed with the values returned from the provider states:
For JSON bodies, use valueFromProviderState.
For headers, use headerFromProviderState.
For query parameters, use queryParameterFromProviderState.
For paths, use pathFromProviderState.
For example, assume that an API call is made to get the details of a user by ID. A provider state can be defined that specifies that the user must be exist, but the ID will be created when the user is created. So we can then define an expression for the path where the ID will be replaced with the value returned from the provider state callback.
.pathFromProviderState("/api/users/${id}", "/api/users/100")You can also just use the key instead of an expression:
.valueFromProviderState('userId', 'userId', 100) // will look value using userId as the keyThis is an extension for the pact DSL. The difference between the default pact DSL and this lambda DSL is, as the name suggests, the usage of lambdas. The use of lambdas makes the code much cleaner.
The lambda DSL solves the following two main issues. Both are visible in the following code sample:
new PactDslJsonArray()
.array() # open an array
.stringValue("a1") # choose the method that is valid for arrays
.stringValue("a2") # choose the method that is valid for arrays
.closeArray() # close the array
.array() # open an array
.numberValue(1) # choose the method that is valid for arrays
.numberValue(2) # choose the method that is valid for arrays
.closeArray() # close the array
.array() # open an array
.object() # now we work with an object
.stringValue("foo", "Foo") # choose the method that is valid for objects
.closeObject() # close the object and we're back in the array
.closeArray() # close the arrayMethods may only be called in certain states. For example object() may only be called when you're currently working on an array whereas object(name)
is only allowed to be called when working on an object. But both of the methods are available. You'll find out at runtime if you're using the correct method.
Finally, the need for opening and closing objects and arrays makes usage cumbersome.
The lambda DSL has no ambiguous methods and there's no need to close objects and arrays as all the work on such an object is wrapped in a lamda call.
When formatting your source code with an IDE the code becomes hard to read as there's no indentation possible. Of course, you could do it by hand but we want auto formatting! Auto formatting works great for the new DSL!
array.object((o) -> {
o.stringValue("foo", "Foo"); # an attribute
o.stringValue("bar", "Bar"); # an attribute
o.object("tar", (tarObject) -> { # an attribute with a nested object
tarObject.stringValue("a", "A"); # attribute of the nested object
tarObject.stringValue("b", "B"); # attribute of the nested object
})
});Start with a static import of LambdaDsl. This class contains factory methods for the lambda dsl extension.
When you come accross the body() method of PactDslWithProvider builder start using the new extensions.
The call to LambdaDsl replaces the call to instance new PactDslJsonArray() and new PactDslJsonBody() of the pact library.
io.pactfoundation.consumer.dsl.LambdaDsl.*import static io.pactfoundation.consumer.dsl.LambdaDsl.newJsonArray;
...
PactDslWithProvider builder = ...
builder.given("some state")
.uponReceiving("a request")
.path("/my-app/my-service")
.method("GET")
.willRespondWith()
.status(200)
.body(newJsonArray((a) -> {
a.stringValue("a1");
a.stringValue("a2");
}).build());import static io.pactfoundation.consumer.dsl.LambdaDsl.newJsonBody;
...
PactDslWithProvider builder = ...
builder.given("some state")
.uponReceiving("a request")
.path("/my-app/my-service")
.method("GET")
.willRespondWith()
.status(200)
.body(newJsonBody((o) -> {
o.stringValue("foo", "Foo");
o.stringValue("bar", "Bar");
}).build());When creating simple json structures the difference between the two approaches isn't big.
{
"bar": "Bar",
"foo": "Foo"
}new PactDslJsonBody()
.stringValue("foo", "Foo")
.stringValue("bar", "Bar")newJsonBody((o) -> {
o.stringValue("foo", "Foo");
o.stringValue("bar", "Bar");
}).build();When we come to more complex constructs with arrays and nested objects the beauty of lambdas become visible!
[
["a1", "a2"],
[1, 2],
[{"foo": "Foo"}]
]new PactDslJsonArray()
.array()
.stringValue("a1")
.stringValue("a2")
.closeArray()
.array()
.numberValue(1)
.numberValue(2)
.closeArray()
.array()
.object()
.stringValue("foo", "Foo")
.closeObject()
.closeArray();newJsonArray((rootArray) -> {
rootArray.array((a) -> a.stringValue("a1").stringValue("a2"));
rootArray.array((a) -> a.numberValue(1).numberValue(2));
rootArray.array((a) -> a.object((o) -> o.stringValue("foo", "Foo")));
}).build();newJsonArray {
newArray {
stringValue("a1")
stringValue("a2")
}
newArray {
numberValue(1)
numberValue(2)
}
newArray {
newObject { stringValue("foo", "Foo") }
}
}As each test will get a new mock server, connections can not be persisted between tests. HTTP clients can cache connections with HTTP/1.1, and this can cause subsequent tests to fail. See #342 and #1383.
One option (if the HTTP client supports it, Apache HTTP Client does) is to set the system property http.keepAlive to false in
the test JVM. The other option is to set pact.mockserver.addCloseHeader to true to force the mock server to
send a Connection: close header with every response (supported with Pact-JVM 4.2.7+).