python / python/mypy

Mypy errors on code that's trivial syntactic transformation of code that it does not error on

Open
#17,715 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug
Dominant language
Python
Stars
20.6k
Forks
3.3k
PR merge metrics
PR metrics pending

Description

Bug Report

Mypy reports a type error for the eta expansion of code that it does not error on. See below for an example.

To Reproduce

from __future__ import annotations

from collections.abc import Callable
from typing import Generic, NamedTuple, TypeVar, TypeVarTuple

DType = TypeVar("DType")
DType2 = TypeVar("DType2")
Shape = TypeVarTuple("Shape")
Shape2 = TypeVarTuple("Shape2")
Dim1 = TypeVar("Dim1")
SeqLen = TypeVar("SeqLen")
BatchLen = TypeVar("BatchLen")


class ndarray(Generic[*Shape, DType]): ...  # noqa: N801


def vmap(
    fun: Callable[[ndarray[*Shape, DType]], ndarray[*Shape2, DType2]],
) -> Callable[[ndarray[Dim1, *Shape, DType]], ndarray[Dim1, *Shape2, DType2]]:
    raise NotImplementedError


def fn(tkns: ndarray[BatchLen, SeqLen, float]):
    _ = vmap(call)(tkns)
    _ = vmap(lambda x: call(x))(tkns)


def call(input_: ndarray[SeqLen, float]) -> ndarray[SeqLen, float]:
    raise NotImplementedError

Expected Behavior

Eta expansion (going from something like vmap(fn) to vmap(lambda x: fn(x)) should always be behavior preserving and also type preserving.

Actual Behavior

Instead Mypy approves of the first call but not the second.

main.py:26: error: Argument 1 to "call" has incompatible type "ndarray[*Shape, DType]"; expected "ndarray[Never, float]"  [arg-type]

Your Environment

This is using the default setting on https://mypy-play.net/?mypy=latest&python=3.12 with Mypy at 1.11.2.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the example from main.py using the default mypy settings with Python 3.12, as demonstrated on mypy-play with mypy 1.11.2. Trace how mypy infers the direct call versus the eta-expanded lambda, and add or update coverage so both equivalent forms receive consistent type-checking results.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
devtools
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
39/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.