docvet 1.13: One Config Line Unlocks Sphinx/RST Support — NumPy Works Out of the Box

[Alberto-Codes]

If your team writes Sphinx-style docstrings — :param name:, :returns:, :raises ValueError: — docvet couldn't check them before. As of v1.13.0, one line in your pyproject.toml changes that. And if you use NumPy-style sections, they already work with zero configuration.

Sphinx/RST support

Add one line to your project config:

[tool.docvet]
docstring-style = "sphinx"

docvet maps Sphinx field-list directives (:param:, :type:, :returns:, :rtype:, :raises:, :ivar:, :cvar:) to the same internal section model used for Google-style, so all existing enrichment rules apply: missing parameters, undocumented exceptions, absent examples — all of it.

def connect(host: str, port: int = 5432) -> Connection:
    """Open a database connection.

    :param host: Hostname or IP address.
    :param port: Port number.
    :returns: An active connection object.
    :raises ConnectionError: If the host is unreachable.
    """

Five rules that have no RST equivalent (require_yields, require_receives, require_warns, require_other_parameters, prefer_fenced_code_blocks) are auto-disabled in Sphinx mode. Override any of them by setting them explicitly in [tool.docvet.enrichment].

NumPy section recognition

NumPy-style underlined headers work automatically in the default Google mode — no configuration change needed:

def transform(data, axis=0):
    """Apply transformation along an axis.

    Parameters
    ----------
    data : array_like
        Input data.

    Returns
    -------
    result : ndarray
        Transformed data.
    """

docvet recognizes these headers and maps them to the internal section model. Same rules, different syntax.

New rules

missing-returns — catches functions that return a value but don't document it. Smart enough to skip stubs, __init__ methods, properties, and simple re-raises. Works in both Google and Sphinx modes.

overload-has-docstring — flags @typing.overload signatures missing docstrings, while the existing missing-docstring rule skips overloads to avoid double-reporting.

By the numbers

• 22 rules across 4 check modules (up from 20)
• 1,424 tests passing (up from 1,350)
• 0 debug sessions across all 4 stories — zero-debug streak at 37 consecutive stories

Try it

pip install --upgrade docvet
docvet check --all

Full changelog · Docs