我收到此错误：[Errno 21] 是一个目录：

Question

# Save directory path in 'path'
path = r'---path '

# Declare a dummy Numpy array (row vector)
result_array = np.empty([1,54])

# Create a list of audio file names 'file_list'
file_list = os.listdir(path)

i=0

for filename in file_list:
    
    # Read WAV file. 'rosa.core.load' returns sampling frequency in 'fs' and audio signal in 'sig'
    sig, fs = rosa.core.load(path + '\\' + file_list[i], sr=None)
    
    # Calculate the average mfcc (utterance-level features) using 'rosa.feat.mfcc()' and 'np.mean' method. '.T' transposes the rows and columns. 'axis=0' indicates average is calculated column-wise
    avg_mfcc_feat = np.mean(rosa.feature.mfcc(y=sig, sr=fs, n_mfcc=26).T,axis=0)
    
    # Calculate the standard deviation of mfcc (utterance-level features) using 'rosa.feat.mfcc()' and 'np.std' method. '.T' transposes the rows and columns. 'axis=0' indicates average is calculated column-wise
    std_mfcc_feat = np.std(rosa.feature.mfcc(y=sig, sr=fs, n_mfcc=26).T,axis=0)
    
    # Calculate the average zero crossing rate (utterance-level feature) using 'rosa.feat.zero_crossing_rate()' and 'np.mean' method. '.T' transposes the rows and columns. 'axis=0' indicates average is calculated column-wise
    zcross_feat = rosa.feature.zero_crossing_rate(sig)
    avg_zcross_feat = np.mean(rosa.feature.zero_crossing_rate(y=sig).T,axis=0)
    
    # Append the three 1D arrays into a single 1D array called 'feat'.
    feat0 = np.append(avg_mfcc_feat, std_mfcc_feat, axis=0)
    
    feat1 = np.append(feat0, avg_zcross_feat, axis=0)
    
    # Save emotion label from file name. 'path' contains directory's address, 'file_list' contains file name, and '\\' joins the two to form file's address
    label = os.path.splitext(os.path.basename(path + '\\' + file_list[i]))[0].split('-')[2]
    
    # Create a new Numpy array 'sample' to store features along with label
    sample = np.insert(feat1, obj=53, values=label)
    
    result_array = np.append(result_array, sample)
    
    i+=1

# Print out the 1D Numpy array
result_array

我在尝试读取内容并循环遍历 Ravdess 数据集的子目录时遇到此错误。有人可以告诉我我做错了什么吗？任何帮助都深表感谢。谢谢！

Answer 1

1. os.listdir() 列出目录的条目，即它包含的文件和目录。如果要递归包含path 下的所有文件和文件夹，请使用os.walk()。

要仅使用来自os.listdir() 的文件，然后使用os.path.isfile() 添加检查：

如果 path 是 existing 常规文件，则返回 True。这遵循符号链接，因此islink() 和isfile() 对于同一路径都可以为真。

for filename in file_list:
    if not os.path.isfile(os.path.join(path, filename)):
        # not a file, skip it
        continue
    # rest of your code here
    sig, fs = rosa.core.load(...)
    ...

2.在for-loop 中，您有两种情况可以使用path + '\\' + file_list[i] - 这应该只是path + '\\' + filename，或者更好的是os.path.join(path, filename)。

• 如果您想要使用文件名迭代的索引，请使用for i, filename in enumerate(file_list): 而不是手动递增i。
• 而不是多次重复os.path.join(path, filename)，而是将它放在循环开始的变量中：

  for filename in file_list:
      filepath = os.path.join(path, filename)
      if not os.path.isfile(filepath):
          continue
  
      sig, fs = rosa.core.load(filepath, sr=None)
      ...