Skip to content

Latest commit

 

History

History
176 lines (131 loc) · 6.24 KB

README.md

File metadata and controls

176 lines (131 loc) · 6.24 KB

openapi-bundler

A utility that merges multiple OpenAPI files into a single file with all external references resolved to local component references.

Why this OpenAPI Bundler

OpenAPI is becoming the de-facto REST API specification, and a lot people have adopted it to write their RESTful API spec.

For small APIs, it is acceptable to have a single OpenAPI definition file, however for a large scale API project, it is hard to manage multiple API specifications together.

For an organization with too many APIs, it is very natural to define sharable object definitions in separate files in order to avoid duplications and foster consistency across an organization. With multiple files inter-connected together to form one API specification, the directory structure and external references are too complicated to manage.

Luckily, there are a lot of tools like editors, parsers and bundlers in the market that support multiple files. These tools can bundle multiple YAML files together to create a final version in JSON or YAML format and de-reference external dependencies.

As our light-rest-4j framework encourages a contract driven design approach, the specification should be created before coding starts. Actually, if the spec is ready, you can use light-codegen to generate the project.

The generator for light-rest-4j generates also the object model (POJO) from the openapi.json where objects are defined in the specification. In addition, light-rest-4j requires the final version of the openapi.json to be included into the service code to do runtime schema validation as well as OAuth 2.0 scope verification at runtime.

This requires that the final version of openapi.json is self-contained and all models should be defined in the components section instead of being de-referenced inline.

Existing bundlers such as swagger-cli cannot handle our specification files correctly, therefore the need arises to create a generally re-usable bundler for OpenAPI definitions.

The tool validates bundled API specifications and offers also independent validation functionality.

Features

Remote Reference

This bundler can resolve all remote references in openapi.yaml which is the main file to be processed. If the reference is an object, it will resolve its internal references first and then move it into components in the generated openapi.json. At the same time the external reference in openapi.yaml will be changed to local reference with #/components/{key}

If the remote reference is not an object, it will be resolved inline in the generated openapi.json file.

If you have separate reference files, you must place these files into the same folder your openapi.yaml, or relative to that folder. For example, if you have common folder that contains all the common OpenAPI files, you might need to copy the common folder into your folder that contains openapi.yaml for your API.

Local Reference

If the reference is an object in components, it will resolve all the remote references in definitions.

If the reference is not an object, an error will occur and the process will exit.

Usage

The bundler assumes that the input file is openapi.yaml and all the remote reference files are in the right path.

Use it as Java utility

java -jar target/openapi-bundler.jar <operation> <arguments....>

Operations are:
  bundle
  validate

Arguments are:
  --dir, -d : The input directory where the YAML files can be found for bundling | validation. Mandatory parameter
  --file, -f : The name of the YAML file to be bundled or validated. Default: openapi.yaml

  --outputFile, -of: The name of the bundled and validated OpenAPI file. Default: openapi.bundled
  --outputDir, -od : The output directory of the bundled and validated file. Default: same as input directory specified in <dir>
  --outputFormat, -o : TThe output format for the bundled file: YAML | JSON | both. Default: YAML

  -debug : to view debug output

# General usage:
  Bundle: 
    java -jar openapi-bundler.jar bundle -d <myFolder> -f <input file> -o <json|yaml|both> -of <output file name> -od <output folder>

  Validate:
    java -jar openapi-bundler.jar bundle -d <myFolder> -f <input file>

# To view debug messages during the bundling or valdiation process, use the utility with -debug
  java -debug -jar openapi-bundler.jar bundle -d <myFolder> -f <input file> -o <json|yaml|both> 

# To view debug messages during the bundling or valdiation process, use the utility with -debug
  java -debug -jar openapi-bundler.jar validate -d <myFolder> -f <input file> 

# Simplified bundler call, with default values, uses openapi.yaml as input and json as output format
java -jar openapi-bundler.jar  bundle -d <myFolder>


To view usage help you can use the following command:

Command:
java -jar openapi-bundler.jar  -h

Usage: [options] operation: The operation to be performed. 
       Supported operations: bundle |validate. Must be specified

  Options:
    --dir, -d
      The input directory where the YAML files can be found. Must be specified
    --file, -f
      The name of the YAML file to be bundled. Default = openapi.yaml
    --output, -o
      The output format for the bundled file: yaml | json | both. Default = json
      Default: json
    --help, -h      
    -debug
      Debug mode
      Default: false

Sample files

openapi.yaml

openapi: 3.0.0

paths: 
  /test:
    get:
      operationId: test
      responses:
        '200':
          content:
            application/json:
              schema:
                type: array
                items:
                  $ref: 'common.yaml#/MyObj'

common.yaml

MyObj:
  type: object
  properties:
    testCode:
      type: string
      readOnly: true
    name:
      type: string
      readOnly: true

Use it in an IDE

Another way to run the bundler is from an IDE. Just set the folder of the openapi.yaml file as a program argument and you can easily debug into it.

Use Docker

There is a Docker image that is published to Docker Hub.

Here is the command line to call it.

TBD

With above command line, you can easily build a script to call it as part of your DevOps flow.