0

我正在尝试将 Matlab 代码转换为 Python,但是当我转换以下行时它会出错:

Matlab代码

md(1,index) = (-1)^bits(m);

Python 等价物

md[index]=(-1)**bits(m)

错误

md[index]=(-1)**bits(m)
TypeError: 'numpy.ndarray' object is not callable

Matlab代码

fdel=2;
fn=10;

zeta=1/sqrt(2);
spb=100;
npts=2000;
fs=2000;
freq_in = zeros(1,2000);      % initialize input frequency array
phi_in = zeros(1,2000);       % initialize input phase array
t = (0:(npts-1))/fs;            % generate vector of sample times
nsettle = fix(npts/10);         % set settle time as 0.1*npts
tsettle = nsettle/fs;           % set settle time
% 
% %       The following three lines of code generate the arrays of the
% %       input frequency and input phase. 
% 
phin1 = 2*pi*fdel*(t-tsettle);
freq_in = [zeros(1,nsettle),fdel*ones(1,npts-nsettle)];
phi_in = [zeros(1,nsettle),phin1(1,(nsettle+1):npts)];
% 
% %       Generate the QPSK input signal and input signal.
% 
nbits = npts/spb;               % Determine number of bits
md = zeros(1,nbits*spb);
bits = round(rand(1,nbits));
for m=1:nbits

 for n=1:spb
  index = (m-1)*spb + n;
  % error making line
    md(1,index) = (-1)^bits(m);
 end
 end

Python代码

fdel=2
fn=10
zeta=1/sqrt(2)
spb=100
npts=2000
fs=2000
freq_in=zeros(2000)
phi_in=zeros(2000)
t=(arange(0,npts-1))/fs
nsettle=fix(npts/10)
tsettle=nsettle/fs
phin1=2*pi*fdel*(t-tsettle)
freq_in=array([zeros(nsettle),fdel*ones(npts-nsettle)])
phi_in=array([zeros(nsettle),phin1[nsettle+1:npts]])
nbits=npts/spb

md=zeros(nbits*spb)
bits=around(np.random.uniform((nbits,)))

for m in arange(0,nbits):
    for n in arange(0,spb):
       index=(m-1)*spb+n
       md[index]=(-1)**bits(m)
4

2 回答 2

4

()此错误是因为您使用而不是索引数组[],例如:

np.arange(10)(1)

给出:

TypeError: 'numpy.ndarray' object is not callable

尽管:

np.arange(10)[1]

给出:

1
于 2013-08-12T12:53:58.030 回答
3

正如 Saullo 在他的回答中指出的那样,您没有以正确的方式建立索引,但是您没有正确移植代码。freq_in并且phi_in没有正确定义并且您没有在 中生成随机向量bits,请查看以下代码:

import numpy as np

fdel = 2
fn = 10
zeta = 1 / np.sqrt(2)
spb = 100
npts = 2000
fs = 2000
freq_in = np.zeros((2000))
phi_in = np.zeros((2000))
t = np.arange(0,npts) / fs
nsettle = np.fix(npts/10)
tsettle = nsettle
phin1 = 2 * np.pi * fdel * (t-tsettle)
freq_in = np.hstack((np.zeros(nsettle), fdel * np.ones(npts-nsettle)))
phi_in = np.hstack((np.zeros(nsettle), phin1[int(nsettle):int(npts)]))

nbits = npts / spb

md = np.zeros(nbits * spb);
bits = np.random.randint(2, size=nbits)
for m in np.arange(0,nbits):
    for n in np.arange(0,spb):
       index = (m-1) * spb + n
       md[index]=(-1)**bits[m]
于 2013-08-12T13:07:34.147 回答