jtd-codebuild 0.8.0

Tool for generating language specific schemas and interfaces code from JSON Type Definition IDL files in yaml format. Powered by jtd-codegen.

JSON Type Definition Code Build

jtd-codebuild is a tool for generating language specific schemas and interfaces code from JSON Type Definition IDL files in either yaml or json format.

This tool is built on top of jtd-codegen so check out the documentation if you don't have a clue about JSON Type Definition.

Prerequisites

• jtd-codegen

Installation

pip install jtd-codebuild

Usage

First, you need to create a configuration file named jtd-codebuild.json in the root of your project. (See Configuration)

Then, write JSON type definition IDL files in either yaml or json format in the directory you specified in the configuration file.

After all of that, you can run the command below to generate code.

jtd-codebuild path/to/the/folder/where/jtd-codebuild.json/is/located

Usage Example

You can find the example project here: jtd_codebuild/tests/fixtures/example_project_1

Also, you can get a sense of how this tool works by looking at the test cases: jtd_codebuild/tests/test_example_project_1.py

Required conventions

Configuration

The script will find jtd-codebuild.json which is the configuration file of this tooling.

root-schema-path

The path to the root schema file. Root schema file will be the entry point of the code generation, where every definition files will be merged into.

Defaults to schema.jtd.yaml

definitions-path (alias: definitions-paths)

The path to single or multiple definition files. This directory will be recursively searched for definition files.

Definition file is a file that contains a single or multiple definitions. Checkout the documentation below for more information.

Defaults to []

output-schema-path

The path for the merged schema file converted in json format.

Defaults to gen/schema.jtd.json

includes

Array of JTD package paths to include.

The path should have jtd-codebuild.json file in it so that this tool can find the codebuild configuration.

If you specifiy a package path, it will reference the package's schema definitions when generating schema file you are working on.

Defaults to []

allow-duplicate-defs

Whether to allow duplicate definitions.

This only applies to the definitions that are not in the same jtd codebuild project.

That is, definitions in definitions-path in each project always raises error if there are duplicate definitions.

Defaults to false

targets

Compile targets.

Defaults to []

It's a JSONRecord contains the object having following properties:

• language (string): The language of the target. We essentially inject this value to jtd-codegen as target language option which is provided as a flag which is like --{language}-out. Available languages can be found in the documentation of jtd-codegen. (See: https://jsontypedef.com/)
• path (string): The path to the directory where the generated code will be placed.

targets - Language Specific Options - Python

• use-pydantic (boolean): Whether to use pydantic as a dataclass decorator provider. If this is set to true, the generated code will use pydantic.dataclasses.dataclass as a dataclass decorator so that you can use pydantic's validation features. Otherwise, the generated code will be plain python dataclasses. Defaults to false.
• use-typeddict (boolean): Whether to use TypedDict instead of dataclass. If this is set to true, the generated code will use TypedDict instead of dataclass. This property cannot be set to true if use-pydantic is set to true. Also, subscriptable option will be ignored if this is set to true.
• property-case (enum): The case of the property names. This property is only effective when use-typeddict is set to true. Available values are snake and camel. Defaults to snake.
• subscriptable (boolean): Whether to make the generated class subscriptable. If this is set to true, the generated class will be subscriptable so that you can access the properties of the class like obj["property"]. Otherwise, the generated class will not be subscriptable. Defaults to false.

targets - Language Specific Options - TypeScript

• tsconfig-path (string): The path to the tsconfig file. This will be used to compile typescript code to javascript code and type declarations. If you want to automatically generate plain javascript artifact with type declarations, you should also provide this option.

Configuration Example

Example congfiguration file is provided below. Copy it and modify it to your needs.

{
  "root-schema-path": "schema.jtd.yaml",
  "definitions-path": "definitions",
  "output-schema-path": "gen/schema.jtd.json",
  "includes": ["../path/to/another/folder/that/contains/jtd-codebuild.json"],
  "targets": [
    {
      "language": "python",
      "path": "gen/python"
    },
    {
      "language": "python",
      "path": "gen/python-pydantic",
      "use-pydantic": true,
      "subscriptable": true
    },
    {
      "language": "python",
      "path": "gen/python-typeddict",
      "use-typeddict": true,
      "subscriptable": true
    },
    {
      "language": "typescript",
      "path": "gen/typescript",
      "tsconfig-path": "tsconfig.build.json"
    }
  ]
}

Root Schema File

Root schema file is the entry point of the code generation.

It will be the file where every definition files will be merged into.

If you don't need a root Schema type, you can just create an empty file.

Definition files

Definition files are sharable files of which each of them contains a single or multiple definitions.

Each declared keys as a root key in the definition file will be merged as a key of definitions object in the root schema file, and those symbols can be shared across the other definition files.

For example, let's say you have a definition file whose code is like below.

book:
  properties:
    id:
      type: string
    title:
      type: string

This can be referenced in other definition files like below.

user:
  properties:
    id:
      type: string
    name:
      type: string
    books:
      elements:
        ref: book

This will be merged as a single schema like below.

{
  "definitions": {
    "book": {
      "properties": {
        "id": {
          "type": "string"
        },
        "title": {
          "type": "string"
        }
      }
    },
    "user": {
      "properties": {
        "id": {
          "type": "string"
        },
        "name": {
          "type": "string"
        },
        "books": {
          "elements": {
            "ref": "book"
          }
        }
      }
    }
  }
}

Checkout more about ref if you don't have a clue about it. https://jsontypedef.com/docs/jtd-in-5-minutes/#ref-schemas

Inheritance

jtd-codebuild supports inheritance between type definitions.

You can do this by using extends keyword like below.

Person:
  properties:
    id:
      type: string
    name:
      type: string

Chef:
  extends: Person
  properties:
    restaurant:
      type: string

HeadChef:
  extends: Chef
  properties:
    sousChef:
      ref: Chef

You can also extend multiple types like below.

Person:
  properties:
    id:
      type: string
    name:
      type: string

MyRestaurantMixin:
  properties:
    restaurant:
      type: string

Chef:
  extends: [Person, MyRestaurantMixin]

Manual dependency management

Since IDL files are basically just a bunch of JSON objects, we need to manually manage the dependency between the definition files.

For example, assume you have a folder structure like the below:

definitions
├── book
│   └── book.jtd.yaml
└── user
    └── user.jtd.yaml

And assume that book.jtd.yaml and user.jtd.yaml are the root definition files of each module.

In this case, you need to annotate that user.jtd.yaml depends on book.jtd.yaml like below.

# user.jtd.yaml
#
# Depends on:
#   - book (at ../book/book.jtd.yaml)
user:
  properties:
    id:
      type: string
    name:
      type: string
    books:
      elements:
        ref: book

Contributing

Any contribution is welcome! Check out CONTRIBUTING.md and CODE_OF_CONDUCT.md for more information on how to get started.

License

jtd-codebuild is licensed under a MIT License.