pydantic-msgspec

Generate msgspec Struct types from Pydantic v2 BaseModel classes. Structs use rename="camel" for JSON keys. Models with a single-string Literal role or type field map to msgspec tag_field / tag for tagged unions.

Install

pip install pydantic-msgspec
# or
uv add pydantic-msgspec

Usage

from pydantic import BaseModel
import msgspec
from pydantic_msgspec import msgspec_from_pydantic, pydantic_from_msgspec


class Item(BaseModel):
    name: str
    count: int = 1


MSItem = msgspec_from_pydantic(Item)
raw = msgspec.json.encode(MSItem(name="x", count=2))
model = pydantic_from_msgspec(Item, MSItem(name="x", count=2))
assert model == Item(name="x", count=2)

Full API reference: Read the Docs.

Development

uv sync --all-extras   # dev + docs
uv run pytest --cov=pydantic_msgspec
uv run sphinx-build -W -b html docs docs/_build

For maintainers

PyPI (trusted publishing with GitLab)

1. On PyPI (or TestPyPI), add a pending publisher or open your project → Manage → Publishing → Add a new pending publisher / trusted publisher for GitLab.
2. Set GitLab namespace and project to your path (for example group/pydantic-msgspec), workflow to GitLab CI/CD, and the correct GitLab hostname if not gitlab.com.
3. Tag a release (v0.1.0) so the deploy job in .gitlab-ci.yml runs; the pipeline uses OIDC (id_tokens) and twine without long-lived API tokens.

Adjust Repository / Documentation URLs in pyproject.toml to match your real GitLab and Read the Docs project slug.

Read the Docs

1. Import the GitLab repository at readthedocs.org.
2. Use the included .readthedocs.yaml; set the default branch and Python version as needed.
3. After the first successful build, update PyPI project URLs if the docs URL differs from the placeholder in pyproject.toml.