As a Library¶

from pathlib import Path

from namelint.visitor import lint_source

source = "def get_user(u):\n    return httpx.get(u)\n"
for finding in lint_source(source, Path("x.py")):
    print(finding.code, finding.name, finding.evidence)

lint_module(tree, path) -> list[Finding] takes a parsed ast.Module. Its signature is flake8's plugin contract on purpose, so a flake8 adapter would need an entry point and a wrapper, with no second copy of the checks.

Findings¶

Every Finding is a frozen dataclass with these fields:

Field	Holds
`code`	the check, such as `NAM106`
`line`, `col`	where the function is defined
`name`	the function's name
`owner`	the enclosing class, if any
`message`	what is wrong, naming the evidence
`evidence`	the fact that produced the finding, such as `calls httpx.get`
`hint`	what to do about it

The evidence field makes a finding arguable: "calls httpx.get" is a fact you can dispute, and "performs I/O" is not.

The checks module¶

namelint.checks holds one pure function per code, each taking a function node and a Context. It imports nothing beyond ast and the two records, touches no filesystem, and reads no configuration. The verb table reaches it through Context.

There is no ruff plugin, because ruff has no plugin API.