Numba try: if array.shape[1] - error: tuple index out of range. works without numba, doesn't work with @njit(fastmath=True, nogil=True, cache=True)

Question

Numba 0.53.1，Python3.7.9，Windows10 64 位

这个 doctest 工作正常：

import numpy as np

def example_numba_tri(yp):
    """
    >>> example_numba_tri(np.array([0.1, 0.5, 0.2, 0.3, 0.1, 0.7, 0.6, 0.4, 0.1]))
    array([[0.1, 0.3, 0.6],
           [0.5, 0.1, 0.4],
           [0.2, 0.7, 0.1]])
    """
    try:
        if yp.shape[1] == 3:
            pass
    except:
        yp = yp.reshape(int(len(yp) / 3), -1, order='F')

    return yp

只需添加@njit(fastmath=True, nogil=True, cache=True):

from numba import njit
import numpy as np

@njit(fastmath=True, nogil=True, cache=True)
def example_numba_tri(yp):
    """
    >>> example_numba_tri(np.array([0.1, 0.5, 0.2, 0.3, 0.1, 0.7, 0.6, 0.4, 0.1]))
    array([[0.1, 0.3, 0.6],
           [0.5, 0.1, 0.4],
           [0.2, 0.7, 0.1]])
    """
    try:
        if yp.shape[1] == 3:
            pass
    except:
        yp = yp.reshape(int(len(yp) / 3), -1, order='F')

    return yp

并得到一个错误：

    numba.core.errors.TypingError: Failed in nopython mode pipeline (step: nopython frontend)
    Internal error at <numba.core.typeinfer.StaticGetItemConstraint object at 0x0000020CB55A7608>.
    tuple index out of range
    During: typing of static-get-item at C:/U1/main.py (1675)
    Enable logging at debug level for details.
    
    File "main.py", line 1675:
    def example_numba_tri(yp):
        <source elided>
        try:
            if yp.shape[1] == 3:
            ^

如何解决它以及为什么会发生这种情况？或者这是一个错误？我读了 https://numba.pydata.org/numba-doc/dev/reference/pysupported.html#pysupported-exception-handling ，但似乎我做了那里写的所有事情。

更新：

https://github.com/numba/numba/issues/6872
这对我也很有用https://numba.pydata.org/numba-doc/latest/user/troubleshoot.html#my-code-doesn-t-compile
分析和重构帮助我仅将大多数 cpu 密集部分转换为 numba
numba.pydata.org/numba-doc/dev/reference/numpysupported.html 好像不支持 order (order='F')

Answer 1

请考虑以另一种方式重写您的代码，因为看起来您的 Numba 代码检查输入数据的类型，因此块 if yp.shape[1] == 3: 在编译阶段被检查，这就是为什么 [=12 不处理它的原因=]

请尝试下面的代码，它与您的代码相同，但是没有 order='F' 不想以任何方式使用 Numba。

from numba import njit
import numpy as np


@njit(fastmath=True, nogil=True, cache=True)
def example_numba_tri(yp):
    return yp.reshape(int(len(yp) / 3), -1)


def wrapper_example_numba_tri(yp):
    if len(yp.shape) > 1:
        if yp.shape[1] == 3:
            return yp
    return example_numba_tri(yp)


if name == 'main':
    x = np.array([0.1, 0.5, 0.2, 0.3, 0.1, 0.7, 0.6, 0.4, 0.1])
    wrapper_example_numba_tri(x)

Numba try: if array.shape[1] - error: tuple index out of range. works without numba, doesn't work with @njit(fastmath=True, nogil=True, cache=True)

Numba try: if array.shape[1] - error: tuple index out of range. works without numba, doesn't work with @njit(fastmath=True, nogil=True, cache=True)

python

try-except

numba