CDDL to Python

Generate Python type definitions from CDDL as TypedDict classes or Pydantic models.

cddl2py converts a parsed CDDL schema into Python source code. By default it emits TypedDict-based definitions that work well for static typing. With the --pydantic flag, it emits BaseModel classes instead.

Install

Use the CLI:

npm install cddl2py

Use the programmatic API:

npm install cddl cddl2py

What It Generates

cddl2py currently maps common CDDL constructs into Python-friendly types, including:

• named CDDL assignments to Python aliases or classes
• groups to TypedDict classes
• optional group fields to NotRequired[...]
• arrays to list[...]
• unions to Union[...]
• literals to Literal[...]
• an optional Pydantic mode that emits BaseModel classes and Field(default=...)

It also normalizes names for Python code by turning type names into PascalCase and field names into snake_case.

CLI

The CLI reads a CDDL file and writes generated Python code to stdout, so the normal workflow is to redirect the output into a .py file.

Generate TypedDict output:

npx cddl2py ./path/to/schema.cddl > ./types.py

Generate Pydantic models:

npx cddl2py --pydantic ./path/to/schema.cddl > ./models.py

Show help:

npx cddl2py --help

Programmatic API

The package exports a single transform() function. It accepts the parsed CDDL AST and returns the generated Python source as a string.

import { parse } from 'cddl'
import { transform } from 'cddl2py'

const ast = parse('./schema.cddl')
const python = transform(ast)

console.log(python)

To generate Pydantic models instead:

import { parse } from 'cddl'
import { transform } from 'cddl2py'

const ast = parse('./schema.cddl')
const python = transform(ast, { pydantic: true })

console.log(python)

Example

Input CDDL:

person = {
  name: tstr,
  age: uint,
  ?nickname: tstr,
}

Generated Python (transform(ast)):

from __future__ import annotations

from typing_extensions import NotRequired, TypedDict

class Person(TypedDict):
    name: str
    age: int
    nickname: NotRequired[str]

Generated Python (transform(ast, { pydantic: true })):

from __future__ import annotations

from typing import Optional
from pydantic import BaseModel

class Person(BaseModel):
    name: str
    age: int
    nickname: Optional[str] = None

Notes

• Generated files include a header comment with the cddl2py version used.
• Pydantic output imports from pydantic, so your Python environment should have it installed if you use --pydantic.
• The CLI validates that the input file exists before attempting to parse it.

If you want to contribute fixes or improvements, see the repository contributing guide.