Commit 407e07ec authored by BO ZHANG's avatar BO ZHANG 🏀
Browse files

tweaks

parent fd859e1e
Loading
Loading
Loading
Loading
Loading
+1 −1
Original line number Diff line number Diff line
@@ -21,7 +21,7 @@ Files/directories with asterisks (``*``) marks are optional.
    │   ├── demo.py                           # Python modules
    │   ├── flip_image.py
    │   ├── scratch.py
    │   └── top_level_interface.py            # the top level interface module
    │   └── api.py                            # the top level interface module
    ├── doc                                   # *sphinx-based documentation directory
    │   ├── build
    │   ├── source
+1 −1
Original line number Diff line number Diff line
%% Cell type:markdown id:697384beaffb53a3 tags:

# How to do Python type annotation

%% Cell type:markdown id:4313baabbbf2ecb tags:

## built-in types

%% Cell type:code id:5fbe06f15bbcd630 tags:

``` python
a: int = 8 # 整数型标注
b: float = 1.23 # 浮点型标注
c: bool = True # 布尔型标注
d: tuple = (1, 2) # 元组型标注
e: tuple[int, float] = (1, 2.34) # 限定类型和元素个数的元组标注
f: tuple[int, ...] = (1, 2, 3, 4, 5., "abc") # 限定类型但不定元素个数的元组型标注
f: tuple[int, ...] = (1, 2, 3, 4, 5) # 限定类型但不定元素个数的元组型标注
g: list = [1, 2, 3] # 一般的列表型标注
h: list[int] = [1, 2, 3] # 对列表型来说不需要标元素个数,但可以明确要求元素类型
```

%% Cell type:markdown id:a06987fbe46de9e4 tags:

## `typing` 模块
对于一些内建类型来说是可以直接用内建类型来标注的,例如以下两句是相等的

%% Cell type:code id:c9d78355e0521512 tags:

``` python
l1: list[int] = [1, 2, 3]

from typing import List
l2: List[int] = [1, 2, 3]
```

%% Cell type:markdown id:8cab70562a62c99 tags:

## Union, Optional和Any

- Union表示并集,Union[float, int]表示可以是float也可以是int
- Optional表示可选,Optional[int]表示可以是None也可以是int
- Any表示任意类型

%% Cell type:markdown id:ca8b1d813d84ad8c tags:

## numpy.typing
我们经常会用到numpy.ndarray进行科学计算,可以有以下几种标注方式:

1. 直接用np.ndarray进行标注, 此时不限定类型(dtype)

%% Cell type:code id:6ebd28b4f4c84ab4 tags:

``` python
import numpy as np

x1: np.ndarray = np.arange(10)
```

%% Cell type:markdown id:622cde4d8d3fb34a tags:

2. 限定类型时不能使用np.ndarray而必须用numpy.typing模块

%% Cell type:code id:36d4e0be053bbd1a tags:

``` python
import numpy.typing as npt

x2: npt.NDArray = np.arange(10, dtype=float)
x3: npt.NDArray[np.int_] = np.arange(10, dtype=int)
x4: npt.NDArray[np.float_] = np.arange(10, dtype=float)
x5: npt.NDArray[np.float32] = np.arange(10, dtype=np.float32)
x6: npt.NDArray[np.bool_] = np.ones(10, dtype=bool)
x7: npt.NDArray[np.complex_] = np.ones(10, dtype=np.complex_)
```

%% Cell type:code id:47b6d72d76ee84a1 tags:

``` python
```