stripe-client 2.0.0

Stripe API Client

A web client for interfacing with Stripes REST API. This client consists of files that have been generated automatically from Stripe's OpenAPI Specification using the generation tool openapi-client.

This library uses Vibe.d for JSON serialization/deserialization and HTTP client logic.

Client Code Organization

• stripe.model :: The package stripe.model contains commonly used data types for the Stripe API, which are generated from the #/components/schemas/ portion of the Stripe OpenAPI Specification.
• stripe.service :: The package stripe.service contains client logic to communicate with REST API endpoints. Generally, the module and class names are produced by concatenating the URL path fragments, e.g. /v1/charges => v1_charges_service.d => class V1ChargesService.
• stripe.security :: Credentials needed to securely access an API, such as API tokens, usernames, and passwords are configured using this module.

Using the Stripe Client

Below is an example of how to use the Stripe Client:

import stripe.security : Security;
import stripe.service.v1_charges_service : V1ChargesService;
import vibe.data.json : serializeToJsonString;

// 1. Stripe's OpenAPI specification has two valid security schemes:
//   - HTTP Basic Auth (named "BasicAuth")
//   - HTTP Bearer Auth (named "BearerAuth")
Security.configureBasicAuth(
    "sk_test_51MFbD...vri",  // Username / API key
    "");                     // With Stripe, the password is always blank.

// 2. Service classes are created from valid URL paths + "Service", e.g. "/v1/charges" => "V1ChargesService".
auto service = new V1ChargesService();

// 3. Each endpoint has a "Params" object which covers any path, query-string, header, or cookie parameters.
auto params = new V1ChargesService.GetChargesParams();

// 4. Some requests have a request body, which will be an argument to the method,
// e.g. "postCharges(params, requestBody, handler)".

// 5. Different HTTP status codes can be associated with different data types.
// Create a handler object and add your own delegates that say what to do with each response.
auto handler = new V1ChargesService.GetChargesResponseHandler();

// 6. This handler is for a successful 200 response, there's also a default handler for errors.
handler.handleResponse200 = (V1ChargesService.GetChargesResponseHandler.ChargeList chargeList) {
  // Simply print out our response in JSON format.
  writeln(serializeToJsonString(chargeList));
};

// 7. Now call the desired endpoint and your handler will be invoked depending on the response.
service.getCharges(params, handler);

Using the numbers in the comments in the example above, we can see the following:

1. Stripe requires an authenticated API Token before it can be used. Learn more about signing up for an account and authentication with an API token here: https://stripe.com/docs/api/authentication
2. Service names are derived from the URL path of the endpoint it corresponds to. Many endpoints can be grouped into a single service, provided that they share the same prefix, e.g. /v1/events and /v1/events/{id}.
3. "Parameters" are shared among all methods that have the same path, e.g. GET /v1/events and POST /v1/events. They include path-parameters, querystring-parameters, header-parameters, and cookie-parameters.
4. An HTTP Request Body is typically only sent with POST and PUT requests, and if one exists, it will be provided as an argument to a service method.
5. The Stripe API may have different response types that are associated with different HTTP status codes, e.g. a 200 OK response might contain card data, but a 404 NOT FOUND response might contain an error description. In the ResponseHandler, you specify how to handle each type of response.
6. Here we provide a delegate that specifies what should be done when the response is a 200 OK. We're given a ChargeList, which was deserialized from the response body. Different fields of the handler class indicate what other responses may be returned.
7. When making an actual request, it can have parameters, a request body, and a handler for the response.

Development

Environment

In order the build this library, the following dependencies are needed in addition to a D compiler:

• Docker: Used to run a mock stripe server during testing.

Testing

Removed from the normal "unittest" configuration profile used during compilation of non-release builds, is a separate configuration that can be used for integration tests. The integration tests set an artificial Stripe API key and interact with a Stipe-Mock server.

To execute these integration tests, run the command:

dub test --config=integration

Regenerating the Client Code from the OpenAPI Specification

The OpenAPI specification is downloaded and saved into json/spec3.json. This file is selected and used with the openapi-client program to regenerate the source code residing under source/stripe.

To regenerate this code, run the command:

dub build --config=generate

Performance Testing

The openapi-client can be invoked manually and passed through the tool Valgrind in order to assess its performance.

Invoke Valgrind to inspect function calls with the command:

valgrind --tool=callgrind --collect-jumps=yes --callgrind-out-file=callgrind_out \
  ../openapi-client/target/openapi-client -- --targetDir=source --openApiSpec=json/spec3.json --packageRoot=stripe

Demangle the symbols so that they are more human readable using the command:

ddemangle callgrind_out > callgrind_out.demangle

Finally, the demangled output can be viewed using the tool kcachegrind.

kcachegrind callgrind_out.demangle

At the time of writing (2023-01-12), approximately 65% of the time is spent parsing the JSON itself, while only 9% of the time is spent writing the source code files.

Future Features

1. Greatly expand integration test coverage for many more useful endpoints.