Recursively get information of all files in specified directory and subdirectories (esp. WAV file properties) and write to text file.

audio_files_info.py

import os # walking through file system
import sys # command lines arguments
import wave # for reading .wav files

# Turn number of seconds (int) to '[minutes] min [seconds] s' (string)
def seconds2timestr(time_in_seconds):
    if time_in_seconds is not None: 
        time_in_minutes = int(time_in_seconds/60)
        remaining_seconds = int(time_in_seconds%60)
        return str(time_in_minutes) + "min " + str(remaining_seconds) + "s"
    else:
        return 'N/A'

def list_files(startpath):
    output_str = ''
    total_size_in_MB = [0]
    number_of_files = [0]
    number_of_wav_files = [0]
    previous_level = 0
    for root, dirs, files in os.walk(startpath):
        dirs.sort(key=lambda s: s.lower()) # sort dirs in place to walk through folders alphabetically
        level = root.replace(startpath, '').count(os.sep)

        # if we went up a level again, print accumulated size
        if level < previous_level:
            for l in range(previous_level,level,-1):
                indent = ' ' * 4 * (l)
                output_str +=  indent+45*"-"+'\n' # horizontal line
                previous_root = os.path.dirname(previous_root) # go one level up
                output_str +=  indent+"TOTAL SIZE"+20*" "+str(total_size_in_MB[l])+" MB\t"+\
                                str(number_of_files[l])+" files\t"+str(number_of_wav_files[l])+" WAV files\n"+\
                                "("+previous_root+")\n\n"
                total_size_in_MB[l] = 0 # reset to zero
                number_of_files[l] = 0 # reset to zero
                number_of_wav_files[l] = 0 # reset to zero
        previous_level = level

        indent = ' ' * 4 * (level)
        output_str += ('{}{}/'.format(indent, os.path.basename(root)))+'\n'
        subindent = ' ' * 4 * (level + 1)

        # sort files alphabetically, case insensitive
        files = sorted(files, key=lambda s: s.lower())

        # keep track of accumulated size at the current level
        if  level>=len(total_size_in_MB)-1: # new level, set to zero
            total_size_in_MB.append(0)
            number_of_files.append(0)
            number_of_wav_files.append(0)
        else: # reset for this level
            total_size_in_MB[level+1] = 0 
            number_of_files[level+1] = 0 
            number_of_wav_files[level+1] = 0 

        for f in files:
            # add size to all levels up to this one
            statinfo = os.stat(root+'/'+f)
            file_size = statinfo.st_size
            for l in range(-1,level+1):
                total_size_in_MB[l+1] += file_size/1000000.0
                number_of_files[l+1] += 1

            # print file info
            if f.endswith('.wav') and not f.startswith('.'):
                for l in range(-1,level+1):
                    number_of_wav_files[l+1] += 1
                try:
                    wav_file = wave.open(root+'/'+f, 'r')
                    fs = wav_file.getframerate()
                    bit_depth = 8*wav_file.getsampwidth()
                    n_chan = wav_file.getnchannels()
                    file_length_in_seconds = wav_file.getnframes()/fs
                    output_str +=  '{}{}'.format(subindent, f)+max((30-len(f)),1)*' '+str(file_size/1000000.0)+' MB\t'+\
                        str(fs)+' Hz\t'+str(bit_depth)+' bit\t'+\
                        str(n_chan)+(' channel\t' if n_chan==1 else ' channels\t')+\
                        seconds2timestr(file_length_in_seconds)+'\n'
                except:
                    output_str += "Unexpected error with "+root+'/'+f+'\n'
            elif not f.startswith('.'): # ignore hidden files
                output_str += ('{}{}'.format(subindent, f))+max((30-len(f)),0)*' '+str(file_size/1000000.0)+' MB'+'\n'

        if files: # if there are any files
            output_str +=  subindent+45*"-"+'\n' # horizontal line
            output_str +=  subindent+"TOTAL SIZE"+20*" "+str(total_size_in_MB[level+1])+" MB\t"+\
                                str(number_of_files[level+1])+" files\t"+str(number_of_wav_files[level+1])+" WAV files\n\n"
        previous_root = root
    # go down all levels and print accumulated size for each
    for l in range(level,-1,-1):
        indent = ' ' * 4 * (l)
        output_str +=  indent+45*"-"+'\n' # horizontal line
        previous_root = os.path.dirname(previous_root) # go one level up
        output_str +=  indent+"TOTAL SIZE"+20*" "+str(total_size_in_MB[l])+" MB\t"+\
                                str(number_of_files[l])+" files\t"+str(number_of_wav_files[l])+" WAV files\n"+\
                                "("+previous_root+")\n\n"
        
    return output_str

def main():
    # parse command line options
    if len(sys.argv)==1: # default path
        PATH = "."
    elif len(sys.argv)==2: # specified path
        PATH = sys.argv[1]
    else:
        print "audio_files_info takes at most 1 command line argument\n"+\
               "Use: python audio_files_info.py [folder_path]."
    
    output_str = list_files(PATH)

    with open('testfile.txt', 'w') as f:
        f.write(output_str)

if __name__ == "__main__":
    main()