Matrix Market#

Matrix Market 交换格式是 NIST 描述的一种基于文本的文件格式。Matrix Market 支持用于稀疏矩阵的 坐标格式 和用于密集矩阵的 数组格式。scipy.io 模块提供了 scipy.io.mmread 和 scipy.io.mmwrite 函数，分别用于读写 Matrix Market 格式的数据。这些函数适用于 numpy.ndarray 或 scipy.sparse.coo_matrix 对象，具体取决于数据是 数组 格式还是 坐标 格式。这些函数可以与 NetworkX 的 convert_matrix 模块中的函数结合使用，以 Matrix Market 格式读写图。

示例#

使用 Matrix Market 的数组格式（array format）读写稠密矩阵对应的图

>>> import scipy as sp
>>> import io  # Use BytesIO as a stand-in for a Python file object
>>> fh = io.BytesIO()

>>> G = nx.complete_graph(5)
>>> a = nx.to_numpy_array(G)
>>> print(a)
[[0. 1. 1. 1. 1.]
 [1. 0. 1. 1. 1.]
 [1. 1. 0. 1. 1.]
 [1. 1. 1. 0. 1.]
 [1. 1. 1. 1. 0.]]

>>> # Write to file in Matrix Market array format
>>> sp.io.mmwrite(fh, a)
>>> print(fh.getvalue().decode('utf-8'))  # file contents
%%MatrixMarket matrix array real symmetric
%
5 5
0.0000000000000000e+00
1.0000000000000000e+00
1.0000000000000000e+00
1.0000000000000000e+00
1.0000000000000000e+00
0.0000000000000000e+00
1.0000000000000000e+00
1.0000000000000000e+00
1.0000000000000000e+00
0.0000000000000000e+00
1.0000000000000000e+00
1.0000000000000000e+00
0.0000000000000000e+00
1.0000000000000000e+00
0.0000000000000000e+00

>>> # Read from file
>>> fh.seek(0)
>>> H = nx.from_numpy_array(sp.io.mmread(fh))
>>> H.edges() == G.edges()
True

使用 Matrix Market 的坐标格式（coordinate format）读写稀疏矩阵对应的图

>>> import scipy as sp
>>> import io  # Use BytesIO as a stand-in for a Python file object
>>> fh = io.BytesIO()

>>> G = nx.path_graph(5)
>>> m = nx.to_scipy_sparse_array(G)
>>> print(m)
  (0, 1)        1
  (1, 0)        1
  (1, 2)        1
  (2, 1)        1
  (2, 3)        1
  (3, 2)        1
  (3, 4)        1
  (4, 3)        1

>>> sp.io.mmwrite(fh, m)
>>> print(fh.getvalue().decode('utf-8'))  # file contents
%%MatrixMarket matrix coordinate integer symmetric
%
5 5 4
2 1 1
3 2 1
4 3 1
5 4 1

>>> # Read from file
>>> fh.seek(0)
>>> H = nx.from_scipy_sparse_array(sp.io.mmread(fh))
>>> H.edges() == G.edges()
True
在本页面