当嵌套 numba 调用的关键字参数数量 > 3 时，Numba 调度错误

Question

使用定义函数

*

时会发生此错误。我可以从三个函数定义案例开始，前两个案例通过，第三个案例是对第二个测试案例的较小修改，但失败。也许不支持*，但是这个错误很有趣，想了解原因。

Numba 版本：'0.56.4' Python 版本：'3.9.17'

通过测试 1

import numba as nb
def test_1(a, b, c, d, e, f, g):
    return a + b + c + d + e + f + g
test_1 = nb.njit(test_1)

def test_2(a, b, c, d, e, f, g):
    return test_1(a, b, c, d, e, f, g)
test_2 = nb.njit(test_2)
test_2(1, 2, 3, 4, 5, 6, 7)

通过测试2

import numba as nb
def test_1(a, b, c, *, d, e, f):
    return a + b + c + d + e + f
test_1 = nb.njit(test_1)

def test_2(a, b, c, d, e, f, g):
    return test_1(a, b, c, d, e, f)
test_2 = nb.njit(test_2)
test_2(1, 2, 3, 4, 5, 6, 7)

测试3失败

import numba as nb
def test_1(a, b, c, *, d, e, f, g):
    return a + b + c + d + e + f
test_1 = nb.njit(test_1)

def test_2(a, b, c, d, e, f, g):
    return test_1(a, b, c, d, e, f, g)
test_2 = nb.njit(test_2)
test_2(1, 2, 3, 4, 5, 6, 7)

错误回溯

---------------------------------------------------------------------------
TypingError                               Traceback (most recent call last)
Cell In[10], line 9
      7 test_2 = nb.njit(test_2)
      8 # test_1(1, 2, 3, 4, 5, 6, 7)
----> 9 test_2(1, 2, 3, 4, 5, 6, 7)

File ~/.cache/pypoetry/virtualenvs/indices-ldm-post-process-danSbNDA-py3.9/lib/python3.9/site-packages/numba/core/dispatcher.py:467, in _DispatcherBase._compile_for_args(self, *args, **kws)
    464         msg = (f"{str(e).rstrip()} \n\nThis error may have been caused "
    465                f"by the following argument(s):\n{args_str}\n")
    466         e.patch_message(msg)
--> 467     error_rewrite(e, 'typing')
    468 except errors.UnsupportedError as e:
    469     # Something unsupported is present in the user code, add help info
    470     error_rewrite(e, 'unsupported_error')

File ~/.cache/pypoetry/virtualenvs/indices-ldm-post-process-danSbNDA-py3.9/lib/python3.9/site-packages/numba/core/dispatcher.py:409, in _DispatcherBase._compile_for_args..error_rewrite(e, issue_type)
    407     raise e
    408 else:
--> 409     raise e.with_traceback(None)

TypingError: Failed in nopython mode pipeline (step: nopython frontend)
Internal error at .
tuple index out of range
During: resolving callee type: type(CPUDispatcher())
During: typing of call at /tmp/ipykernel_3435313/3644502072.py (6)

Enable logging at debug level for details.

我在 Numba github

提出了一个问题

Answer 1

该错误似乎与

*args

函数定义中使用

test_1

语法有关。 Numba 不支持此语法。

要修复该错误，您可以使用

**kwargs

语法。以下是如何修改

test_1

和

test_2

函数的示例：

python
import numba as nb

def test_1(a, b, c, **kwargs):
    d, e, f, g = kwargs.values()
    return a + b + c + d + e + f + g

test_1 = nb.njit(test_1)

def test_2(a, b, c, d, e, f, g):
    return test_1(a, b, c, d=d, e=e, f=f, g=g)

test_2 = nb.njit(test_2)
test_2(1, 2, 3, 4, 5, 6, 7)

此代码应该可以正常运行，没有任何错误。