Numpy

1: pip install numpy

import numpy

import numpy as np

 1: import numpy as np
 2: 
 3: # 使用 ndarray
 4: A = np.array([[1, 2], [3, 4]])
 5: B = np.array([[5, 6], [7, 8]])
 6: print('=== A ===')
 7: print(A)
 8: print('=== B ===')
 9: print(B)
10: 
11: print('=== A * B ===')
12: print(A * B)  # 元素對應相乘
13: 
14: # 使用 matrix
15: A_matrix = np.matrix([[1, 2], [3, 4]])
16: B_matrix = np.matrix([[5, 6], [7, 8]])
17: print('=== A_matrix * B_matrix ===')
18: print(A_matrix * B_matrix)  # 矩陣乘法

 1: import numpy as np
 2: 
 3: np4 = np.array( [[1, 2, 4], [3, 4, 5]] )
 4: print("shape:", np4.shape)
 5: print("np4:\n", np4)
 6: print("取出第0列(row):",np4[0])
 7: print("取出第0行(column):",np4[:, 0])
 8: np5 = np.array([np.arange(1, 10, 2), np.arange(10, 20, 2)])
 9: print('np5:\n', np5)
10: np6 = np.arange(8).reshape(2, 4)
11: print('np6:\n', np6)
12: print(np6.ndim, np6.shape, np6.dtype)

1: import numpy as np
2: 
3: np7 = np.arange(24).reshape(2, 3, 4)
4: print('np7:\n',np7)
5: 
6: np8 = np.arange(13, 60, 2).reshape(2, 3, 4)
7: print("np8:\n", np8)
8: print(np8.ndim, np8.shape, np8.dtype)

1: import numpy as np
2: ary = np.random.randint(10, size=(3,4))
3: print(ary)
4: np.save('ndarray', ary)
5: np.savetxt('ndarray.txt', ary)
6: np.savetxt('ndarray.csv', ary, delimiter=',')

1: ls ndarray*
2: cat ndarray.npy
3: cat ndarray.txt
4: cat ndarray.csv

1: import numpy as np
2: ary1 = np.load('ndarray.npy')
3: print(ary1)
4: ary2 = np.loadtxt('ndarray.txt')
5: print(ary2)
6: ary3 = np.loadtxt('ndarray.csv', delimiter=',')
7: print(ary3)

import timeit

print(timeit.timeit(stmt='import numpy as np;x = np.arange(100);x**3'))
print(timeit.timeit(stmt='for i in range(100): a=i**3'))

1: %timeit [i**3 for i in range(1000)]

1: import numpy as np
2: ar = np.arange(1000)
3: %timeit ar**3

1: import numpy as np
2: arr = np.array([10, 20, 30, 40, 50])
3: # 遍歷、輸出每個元素
4: for elem in arr:
5:     print(elem)

1: import numpy as np
2: a = np.arange(12).reshape(3, 4)
3: print(a)
4: print(a[1])
5: print(a[:,1])
6: print(a[1:3, 1:3])

 1: import numpy as np
 2: 
 3: a = np.arange(12).reshape(3, 4)
 4: print(a)
 5: a_delCol1 = np.delete(a, 1, 0)
 6: print('刪除第1行')
 7: print(a_delCol1)
 8: a_delRow1 = np.delete(a, 1, 1)
 9: print('刪除第1列')
10: print(a_delRow1)
11: print('現在的a')
12: print(a)

1: import numpy as np
2: a = np.array([[1, 0],
3:               [2, 3]])
4: print(a)
5: print('--Matrix transpose--')
6: print(a.transpose())

1: import numpy as np
2: 
3: C = np.array([5, -1, 3, 9, 0])
4: print(C<=0)
5: # 將陣列中小於等於0的元素取代為0;其他轉為1
6: C[C<=0] = 0
7: C[C>0] = 1
8: print(C)

 1: import numpy as np
 2: ar = np.arange(1,5)
 3: print(ar.prod())
 4: 
 5: ar1 = np.array([np.arange(1,4),np.arange(4,7),np.arange(7,10)])
 6: print(ar1)
 7: print(np.prod(ar1, axis=1))
 8: print(ar1.sum())
 9: print(ar1.mean())
10: print(np.median(ar1))

1: import numpy as np
2: 
3: A = np.array([[4, -7], [2, -3]])
4: print("A:\n", A)
5: B = np.linalg.inv(A)
6: print("B:\n", B)
7: print("A dot B:\n", A.dot(B))

numpy.sum(a, axis=None, dtype=None, out=None, keepdims=<no value>, initial=<no value>, where=<no value>)

 1: import numpy as np
 2: 
 3: np_array_2x3 = np.array([[0,2,4],[1,3,5]])
 4: print('=====Array=====')
 5: print(np_array_2x3)
 6: print('=====列=====')
 7: print(np.sum(np_array_2x3, axis = 0))
 8: print('=====行=====')
 9: print(np.sum(np_array_2x3, axis = 1))
10: print('=====陣列總和1=====')
11: print(np.sum(np_array_2x3))
12: print('=====陣列總和2=====')
13: print(np_array_2x3.sum())
14: 

1: import numpy as np
2: print('=====一維tile=====')
3: b = np.array([[1, 2], [3, 4]])
4: b1 = np.tile(b, 2)
5: print(b1)
6: print('=====二維tile=====')
7: b2 = np.tile(b, (2, 3))
8: print(b2)

1: numpy.allclose(a, b, rtol=1e-05, atol=1e-08, equal_nan=False)

 1: import numpy as np
 2: import math
 3: 
 4: a = np.array([[2,2,2],
 5:               [3,3,3]])
 6: b = np.array([[2,3,2],
 7:               [3,3,2]])
 8: print("=====allclose(atol=0.5)=====")
 9: print(np.allclose(a, b, atol=0.5))
10: print("=====allclose(atol=1.0)=====")
11: print(np.allclose(a, b, atol=1.0))
12: print("=====equal()=====")
13: print(np.equal(a,b))
14: 
15: a = np.array([[2,2,2],
16:               [3,3,3.1]])
17: print('=====alllose(atol=11.0)=====')
18: print(np.allclose(a, b, atol=1.0))
19: print('=====isclose()=====')
20: print(np.isclose(a, b, atol=1.0))

1: import numpy as np
2: 
3: ar = np.array([[1,2,3],[4,5,6],[2,3,4]])
4: print(ar)
5: print(ar+ar)
6: print(ar**.5)
7: 
8: ar1 = np.array([[2,2],[3,3],[1,1]])
9: print(ar.dot(ar1))  #陣列dot

 1: import numpy as np
 2: 
 3: xr = np.array([1.1, 1.2, 1.3, 1.4, 1.5])
 4: yr = np.array([2.1, 2.2, 2.3, 2.4, 2.5])
 5: cond = np.array([True, False, True, True, False])
 6: 
 7: result = [(x if c else y) for x, y, c in zip(xr, yr, cond)]
 8: print(result)
 9: print(type(result))
10: print(np.where(cond, xr, yr))
11: print(type(result))
12: #結果都傳回list,為何有,的差異

1: import numpy as np
2: 
3: x1 = np.arange(9.0).reshape((3,3))
4: print(x1)
5: x2 = np.arange(1, 4)
6: print(x2)
7: print(np.multiply(x1,x2))

1: import numpy as np
2: a = np.arange(0, 12).reshape(3,4)
3: print(a)
4: np.save('a', a)
5: # 讀入
6: b = np.load('a.npy')
7: print(b)

1: import numpy as np
2: aAry = [1,2,3,4,5,6]
3: bAry = [7,8,9,10,11,12]
4: #save
5: np.savez('ab.npz', a = aAry, b = bAry)
6: #load
7: myZip = np.load('ab.npz')
8: print(myZip['a'])
9: print(myZip['b'])

 1: import numpy as np
 2: import requests
 3: from io import StringIO
 4: 
 5: url = 'https://letranger.github.io/PythonCourse/scores.csv'
 6: response = requests.get(url)
 7: 
 8: data = np.genfromtxt(StringIO(response.text), delimiter=',')
 9: 
10: print(data)

 1: import pandas as pd
 2: import numpy as np
 3: 
 4: # 從網路上讀取 CSV 文件
 5: url = 'https://letranger.github.io/PythonCourse/scores.csv'
 6: dataframe = pd.read_csv(url)
 7: 
 8: numpy_array = dataframe.to_numpy()
 9: 
10: print(numpy_array)

1: import numpy as np
2: 
3: # 建立一個包含 Python 字典的 Numpy 陣列
4: arr = np.array([{'name': 'Alice'}, {'name': 'Bob'}])
5: 
6: # 保存陣列到 'array_with_dict.npy' 文件中，使用 allow_pickle=True
7: np.save('array_with_dict.npy', arr, allow_pickle=True)

1: import numpy as np
2: 
3: def malicious_function():
4:     print("我是惡意代碼，你已經被入侵了啊哈哈哈....")
5: 
6: arr = np.array([malicious_function], dtype=object)
7: 
8: np.save('malicious_array.npy', arr, allow_pickle=True)

1: import numpy as np
2: 
3: # 在載入之前，重新定義該函數
4: def malicious_function():
5:     print("我是比較沒殺傷力的惡意代碼QQ，你已經被入侵了....")
6: 
7: loaded_arr = np.load('malicious_array.npy', allow_pickle=True)
8: 
9: loaded_arr[0]()

 1: import pickle
 2: import os
 3: 
 4: class MaliciousClass:
 5:     def __reduce__(self):
 6:         # 在反序列化時會執行這行，這裡我們使用 os.system() 執行命令
 7:         return (os.system, ("echo '我是惡意代碼，你已經被入侵了啊哈哈哈....'",))
 8: 
 9: with open('malicious.pkl', 'wb') as f:
10:     pickle.dump(MaliciousClass(), f)

1: import pickle
2: 
3: # 從 pickle 檔案中載入資料，會執行 __reduce__ 方法中的惡意代碼
4: with open('malicious.pkl', 'rb') as f:
5:     pickle.load(f)