如果我的稀疏矩阵有多个非零值区域,我试图提取多个子矩阵.
例如,假设我有以下矩阵:
x = np.array([0,0,0,0,0,0],
[0,1,1,0,0,0],
[0,1,1,0,0,1],
[0,0,0,0,1,1],
[0,0,0,0,1,0])
Run Code Online (Sandbox Code Playgroud)
然后我需要能够提取具有非零值的区域,即
x_1 = [[1,1]
[1,1]]
Run Code Online (Sandbox Code Playgroud)
和
x_2 = [[0,1],
[1,1],
[1,0]]
Run Code Online (Sandbox Code Playgroud)
我一直在使用np.where()来查找非零值的索引并仅返回一个子矩阵的区域,但是如何将其扩展到稀疏矩阵中的所有可能的子区域?
谢谢!
程序:
码:
def delrc(arr):
while True: # delete leading rows with all zeros
if np.all(arr[0]==0):
arr=np.delete(arr,0,axis=0)
else: break
while True: # delete trailing rows with all zeros
if np.all(arr[-1]==0):
arr=np.delete(arr,-1,axis=0)
else: break
while True: # delete leading cols with all zeros
if np.all(arr[:,0]==0):
arr=np.delete(arr,0,axis=1)
else: break
while True: # delete trailing cols with all zeros
if np.all(arr[:,-1]==0):
arr=np.delete(arr,-1,axis=1)
else: break
return arr
def rcsplit(arr):
if np.all(arr==0): return [] # if all zeros return
global res
arr = delrc(arr) # delete leading/trailing rows/cols with all zeros
print arr
indr = np.where(np.all(arr==0,axis=1))[0]
indc = np.where(np.all(arr==0,axis=0))[0]
if not indr and not indc: # If no further split possible return
res.append(arr)
return
arr=np.delete(arr,indr,axis=0) #delete empty rows in between non empty rows
arr=np.delete(arr,indc,axis=1) #delete empty cols in between non empty cols
arr=np.split(arr,indc,axis=1) # split on empty (all zeros) cols
print arr
arr2=[]
for i in arr:
z=delrc(i)
arr2.extend(np.split(z,indr,axis=0)) # split on empty (all zeros) rows
for i in arr2:
rcsplit(np.array(i)) # recursive split again no further splitting is possible
if __name__=="__main__":
import numpy as np
res = []
arr = np.array([[0,0,0,0,0,0],
[0,1,1,0,0,0],
[0,1,1,0,0,1],
[0,0,0,0,1,1],
[0,0,0,0,1,0]])
rcsplit(arr)
for i in res: print i
Run Code Online (Sandbox Code Playgroud)
| 归档时间: |
|
| 查看次数: |
453 次 |
| 最近记录: |