Numpy 2d和1d阵列到乳胶bmatrix

时间:2013-06-16 01:00:42

标签: python numpy latex

我正在寻找一种将numpy数组迁移到latex bmatrix的简洁方法。它应该适用于2d阵列和水平和垂直1d阵列。

示例

A = array([[12, 5, 2],
           [20, 4, 8],
           [ 2, 4, 3],
           [ 7, 1,10]])

print A              #2d array
print A[0]           #horizontal array
print A[:,0, None]   #vertical array

array_to_bmatrix(A)
array_to_bmatrix(A[0])
array_to_bmatrix(A[:,0, None])

输出:

[[12  5  2]
 [20  4  8]
 [ 2  4  3]
 [ 7  1 10]]

[12  5  2]

[[12]
 [20]
 [ 2]
 [ 7]]

\begin{bmatrix} 
 12.000 & 5.000 & 2.000 & \\
 20.000 & 4.000 & 8.000 & \\
 2.000 & 4.000 & 3.000 & \\
 7.000 & 1.000 & 10.000 & \\
\end{bmatrix}

\begin{bmatrix} 
 12.000 & 5.000 & 2.000
\end{bmatrix}

\begin{bmatrix} 
 12.000 & \\
 20.000 & \\
 2.000 & \\
 7.000 & \\
\end{bmatrix}

尝试解决方案

def array_to_bmatrix(array):
    begin = '\\begin{bmatrix} \n'
    data = ''
    for line in array:        
        if line.size == 1:
            data = data + ' %.3f &'%line
            data = data + r' \\'
            data = data + '\n'
            continue
        for element in line:
            data = data + ' %.3f &'%element

        data = data + r' \\'
        data = data + '\n'
    end = '\end{bmatrix}'
    print begin + data + end  

此解决方案适用于垂直和二维数组,但它将水平数组输出为垂直数组。

array_to_bmatrix(A[0])

输出:

\begin{bmatrix} 
 12.000 & \\
 5.000 & \\
 2.000 & \\
\end{bmatrix}

9 个答案:

答案 0 :(得分:16)

numpy数组的__str__方法已经为您完成了大部分格式化操作。让我们利用它;

import numpy as np

def bmatrix(a):
    """Returns a LaTeX bmatrix

    :a: numpy array
    :returns: LaTeX bmatrix as a string
    """
    if len(a.shape) > 2:
        raise ValueError('bmatrix can at most display two dimensions')
    lines = str(a).replace('[', '').replace(']', '').splitlines()
    rv = [r'\begin{bmatrix}']
    rv += ['  ' + ' & '.join(l.split()) + r'\\' for l in lines]
    rv +=  [r'\end{bmatrix}']
    return '\n'.join(rv)

A = np.array([[12, 5, 2], [20, 4, 8], [ 2, 4, 3], [ 7, 1, 10]])
print bmatrix(A) + '\n'

B = np.array([[1.2], [3.7], [0.2]])
print bmatrix(B) + '\n'

C = np.array([1.2, 9.3, 0.6, -2.1])
print bmatrix(C) + '\n'

返回:

\begin{bmatrix}
  12 & 5 & 2\\
  20 & 4 & 8\\
  2 & 4 & 3\\
  7 & 1 & 10\\
\end{bmatrix}

\begin{bmatrix}
  1.2\\
  3.7\\
  0.2\\
\end{bmatrix}

\begin{bmatrix}
  1.2 & 9.3 & 0.6 & -2.1\\
\end{bmatrix}

答案 1 :(得分:1)

另一个,灵感来自罗兰史密斯的回答 支持科学记数法

def bmatrix(a):
    """Returns a LaTeX bmatrix

    :a: numpy array
    :returns: LaTeX bmatrix as a string
    """
    if len(a.shape) > 2:
        raise ValueError('bmatrix can at most display two dimensions')
    temp_string = np.array2string(a, formatter={'float_kind':lambda x: "{:.2e}".format(x)})
    lines = temp_string.replace('[', '').replace(']', '').splitlines()
    rv = [r'\begin{bmatrix}']
    rv += ['  ' + ' & '.join(l.split()) + r'\\' for l in lines]
    rv +=  [r'\end{bmatrix}']
    return '\n'.join(rv)

结果:

\begin{bmatrix}
  7.53e-04 & -2.93e-04 & 2.04e-04 & 5.30e-05 & 1.84e-01 & -2.43e-05\\
  -2.93e-04 & 1.19e-01 & 2.96e-01 & 2.19e-01 & 1.98e+01 & 8.61e-03\\
  2.04e-04 & 2.96e-01 & 9.60e-01 & 7.42e-01 & 4.03e+01 & 2.45e-02\\
  5.30e-05 & 2.19e-01 & 7.42e-01 & 6.49e-01 & 2.82e+01 & 1.71e-02\\
  1.84e-01 & 1.98e+01 & 4.03e+01 & 2.82e+01 & 5.75e+03 & 1.61e+00\\
  -2.43e-05 & 8.61e-03 & 2.45e-02 & 1.71e-02 & 1.61e+00 & 7.04e-03\\
\end{bmatrix}

答案 2 :(得分:0)

执行此操作时:

    for line in array:

您正在迭代array的第一维。当数组是1-D时,最终会迭代值。在进行此迭代之前,您需要确保array确实是2-D。一种方法是通过numpy.atleast_2d传递参数:

import numpy as np

def array_to_bmatrix(array):
    array = np.atleast_2d(array)
    begin = '\\begin{bmatrix} \n'
    data = ''
    for line in array:

答案 3 :(得分:0)

我对使用Python的打印输出不满意。矩阵可能太大,导致包裹。 这是用于打印2d矩阵的LaTeX文本的代码。

def bmatrix(a):
    text = r'$\left[\begin{array}{*{'
    text += str(len(a[0]))
    text += r'}c}'
    text += '\n'
    for x in range(len(a)):
        for y in range(len(a[x])):
            text += str(a[x][y])
            text += r' & '
        text = text[:-2]
        text += r'\\'
        text += '\n'
    text += r'\end{array}\right]$'

    print text

这给出了这个

$\left[\begin{array}{*{16}c}
2.01 & 0 & 0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 \\
0 & 2.01 & 0 & 0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 \\
0 & 0 & 2.01 & 0 & 0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 \\
-1 & 0 & 0 & 2.01 & 0 & 0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & -1 & 0 & 0 & 2.01 & 0 & 0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & -1 & 0 & 0 & 2.01 & 0 & 0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & -1 & 0 & 0 & 2.01 & 0 & 0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & -1 & 0 & 0 & 2.01 & 0 & 0 & -1 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 & 2.01 & 0 & 0 & -1 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 & 2.01 & 0 & 0 & -1 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 & 2.01 & 0 & 0 & -1 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 & 2.01 & 0 & 0 & -1 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 & 2.01 & 0 & 0 & -1 \\
-1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 & 2.01 & 0 & 0 \\
0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 & 2.01 & 0 \\
0 & 0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 & 2.01 \\
\end{array}\right]$

答案 4 :(得分:0)

我已经尝试了一个全面的解决方案,以便个人甚至不需要编写一个最小的脚本就可以完成它。我为浮动,格式化,复杂和Pandas数组增加了灵活性。请使用(array_to_latex)[https://pypi.org/project/array-to-latex/]并提供反馈。

答案 5 :(得分:0)

此外,除了前面的答案外,您还可以通过这种方式从阵列中生成乳胶

from IPython.display import *
from numpy import *
A = array([[12, 5, 2],
           [20, 4, 8],
           [ 2, 4, 3],
           [ 7, 1,10]])
list=A

str1 ='$$' +'\\begin{bmatrix}'+ '&\\\\'.join(str(e) for e in list)+ '\\end{bmatrix}'+'$$'
print(str1 )
str1 = str1.replace('[', ' ')
str1 = str1.replace(']', ' ')

display(Latex(str1))

答案 6 :(得分:0)

另一个选择是使用sympy:首先将数组转换为sympy.Matrix,然后使用sympy.latex函数。

答案 7 :(得分:0)

进一步的答案,灵感来自罗兰·史密斯(Roland Smith)的答案:

def matToTex(a, roundn=2, matrixType = "b",rowVector = False):
    if type(a) != np.ndarray:
        raise ValueError("Input must be np array")
    if len(a.shape) > 2:
        raise ValueError("matrix can at most display two dimensions")
    if matrixType not in ["b","p"]:
        raise ValueError("matrix can be either type \"b\" or type \"p\"")
    if rowVector:
        if not (len(a.shape) != 1 or a.shape[0] != 1):
            raise ValueError("Cannot rowVector this bad boi, it is not a vector!")
    lines = str(a).splitlines()
    ret = "\n\\begin{"+matrixType+"matrix}\n"
    for line in lines:
        line = re.sub("\s+",",",re.sub("\[|\]","",line).strip())
        nums = line.split(",");
        if roundn != -1:
            nums = [str(round(float(num),roundn)) for num in nums]
        if rowVector:
            ret += " \\\\\n".join(nums)
        else:
            ret += " & ".join(nums)+" \\\\ \n"
    ret += "\n\\end{"+matrixType+"matrix}\n"
    ret = re.sub("(\-){0,1}0.[0]* ","0 ",ret)
    print(ret)

答案 8 :(得分:-1)

尝试array_to_latex (pip install)。我正是出于这个原因写的。请提供您的反馈意见。

它具有默认值,但还允许您自定义格式(指数,小数位数)并处理复数,并且可以将结果“弹出”到剪贴板中(无需复制转储到屏幕上的文本)。

github存储库中的一些示例。 https://github.com/josephcslater/array_to_latex