Python报错：'numpy.ndarray' object is not callable 排查求助

问题描述 #

代码此前运行正常，今日首次全量运行无问题，但未做任何修改二次运行时，所有调用Phase_Extractor函数的单元格均触发TypeError: 'numpy.ndarray' object is not callable错误，怀疑与[]和()误用有关但无法定位。

代码片段 #

def Phase_Extractor(fpath):

    n = 1 #air refractive index
    wvl = 632E-9

    displacement = []
    phase = []

    tdms_file1 = TdmsFile.read(fpath)
    for group in tdms_file1.groups():
        groupname = group.name
        for channel in group.channels():
            channel_name = channel.name
            # Access dictionary of properties:
            properties = channel.properties
            # Access numpy array of data for channel:
            velocity_volts1 = channel[:]
            # Access a subset of data
            data_subset = channel[0:100]
            
    velocity = velocity_volts1/1000 ########## Convert from volts to m/s ################################

    time = np.arange(0,5,0.000005) 

    displacement = scipy.integrate.cumtrapz(velocity, time) ################ convert to displacement from velocity #########

    for i in range(0, len(displacement)):
        phase.append(((2*np.pi*n*displacement[i]))/wvl) ############# Convert to phase ################################
        
    phase = np.append(phase, phase[0-1]) ################### make phase list same size as time step list ###################  
        

    return phase

报错栈 #

TypeError                                 Traceback (most recent call last)
Input In [38], in <cell line: 1>()
----> 1 MDP500Full = MinPhaseFinder(r"\Users\seanl\Documents\Master Thesis\500HzFullvol Converted.tdms", 500)
  2 MDP500Half = MinPhaseFinder(r"\Users\seanl\Documents\Master Thesis\500HzHalfvol Converted.tdms", 500)
  3 MDP500Quarter = MinPhaseFinder(r"\Users\seanl\Documents\Master Thesis\500HzQuartervol Converted.tdms", 500)

Input In [26], in MinPhaseFinder(fpath, sf)
      1 def MinPhaseFinder(fpath, sf):
----> 3     filtered_phase = butter_highpass_filter(Phase_Extractor(fpath), 7 , 20000)
      5     X = fft(filtered_phase)
      6     sr = 200E3

Input In [11], in Phase_Extractor(fpath)
     26 time = np.arange(0,5,0.000005) 
     28 displacement = scipy.integrate.cumtrapz(velocity, time) 
---> 30 for i in range(0, len(displacement)):
     31     phase.append(((2*np.pi*n*displacement[i]))/wvl) 
     33 phase = np.append(phase, phase[0-1]) 

TypeError: 'numpy.ndarray' object is not callable

原因与解决方法 #

核心原因 #

报错出现在len(displacement)这一步，说明内置函数len被意外覆盖成了numpy数组（比如之前运行过len = np.array(...)这类代码）。Jupyter Notebook的全局变量会跨单元格保留，首次全量运行时未污染len，二次运行时因之前的全局变量残留，导致调用len()时把数组当成函数执行，触发错误。

解决步骤 #

1. 重启Jupyter内核：清除所有被污染的全局变量，重新运行全部代码即可恢复正常。
2. 检查变量命名：排查所有代码单元格，确保没有使用len、range、append等内置函数名作为变量名，避免再次污染。

代码优化建议 #

• 将函数内的常量（n、wvl）放在函数内部，避免全局变量意外修改。
• 遍历TDMS文件的channel时，当前代码会覆盖velocity_volts1，仅保留最后一个channel的数据，若需处理多channel需调整逻辑。
• 用向量化操作替代for循环提升效率：

  # 替换原for循环
  phase = (2 * np.pi * n * displacement) / wvl
  

• 简化末尾的数组追加操作：

  phase = np.append(phase, phase[-1])  # 替代phase[0-1]，更易读
  

内容的提问来源于stack exchange，提问作者Sean Lane