RESTinstance

Robot Framework test library for (RESTful) JSON APIs

Why?

1. RESTinstance relies on Robot Framework's language-agnostic, natural language syntax for API tests. It is not tied to any particular programming language nor development framework. Using RESTinstance requires little, if any, programming knowledge. It builts on long-term technologies with well established communities, such as HTTP, JSON (Schema), OpenAPI and Robot Framework.

2. It validates JSON using JSON Schema, guiding you to write API tests to base on constraints rather than on specific values (e.g. "email must be valid" vs "email is foo@bar.com"). This approach reduces test maintenance when the values responded by the API are prone to change. Although values are not required, you can still test whenever they make sense (e.g. GET response body from one endpoint, then POST some of its values to another endpoint).

3. It generates JSON Schema for requests and responses automatically, and the schema gets more accurate by you writing tests. The schema is a contract between different teams, or functions (backend, frontend, partners), to agree on what kind of data the API handles. Additionally, you can mark validations to be skipped and rather use the tests to define how the API should work - then the schema also acts as a design. The schema can be further extended to an OpenAPI specification (manually for now, generating also this is planned), which RESTinstance can also test requests and responses against. This leads to very clean looking tests.

Installation

Python (3 only):

pip install --upgrade RESTinstance

The package is hosted in PyPi.

Using Docker

If you have Docker available, prefer rfdocker and just add RESTinstance to your requirements.txt.

Alternatively, a ready image is available at DockerHub.

Usage

See keyword documentation.

The most common use cases for RESTinstances are:

1. Testing for JSON types and formats using JSON Schema validations. For examples, see tests/validations.robot.

2. Flow-driven API tests, i.e. multiple APIs are called for the result. For examples, see tests/methods.robot.

3. Testing API requests and responses against a schema or a specification. For examples testing against JSON schemas, see tests/schema.robot and for testing against Swagger 2.0 specification, see tests/spec.robot.

(TODO: embed examples here)

Development

The issues and requests are tracked in GitHub. We kindly do take pull requests (please mention if you do not want to be listed as contributors).

Running tests

Docker is mandatory for running the library's own tests:

./test

System under test

The test API is implemented by mounterest, which in turn builds on mountebank.

In the scope of library's tests, mounterest acts as a HTTP proxy to Typicode's live JSON server and uses mountebank's injections to enrich responses slightly, so that they better match to this library's testing needs. Particularly, it allows us to test the library with non-safe HTTP methods (POST, PUT, PATCH, DELETE) by mimicking their changes in the state only, instead of trying to issue them on the live server. The state is cleared between the test runs.

Credits

RESTinstance is licensed under Apache License 2.0.

RESTinstance was originally written by Anssi Syrjäsalo, and was initially presented at RoboCon 2018.

Python libraries

We use the following Python libraries and tools under the hood:

• GenSON, by Jon "wolverdude" Wolverton, for JSON Schema generation
• Flex, by Piper Merriam, for Swagger 2.0 validation
• jsonschema, by Julian Berman, for JSON Schema draft-04 validation
• requests, by Kenneth Reitz, for Python HTTP requests

See requirements.txt for a full list of all the (direct) dependencies.