Udacity Nanotrading Transcript Cleaner

cleansrt.py

"""
Turn the .srt files into .txt files
1. unzip the lesson files into a directory
2. run this code in that directory.
the code will strip out the timestamps and carriage returns
and produce a text file that you can use for reference and class notes.
"""
import re
from os import listdir, getcwd
from os.path import isfile, join

currpath = getcwd()
thefiles = [f for f in listdir(currpath) if isfile(join(currpath, f))]

for f in thefiles:
    file = open(f, mode='r', encoding = "ISO-8859-1")
    # read it all
    raw_text = file.read()
    raw_text = re.sub('^\d.*?$', '', raw_text, flags=re.M|re.S)
    raw_text = re.sub('\n\n\n\n', ' ', raw_text, flags=re.M|re.S)
    fixed = open(f.replace('.srt', '.txt') , mode='w')
    fixed.write(raw_text)
    
print('done')