Lecture Notes on 02 November 2009

# This program prompts the use to enter the name of a text file.
# It then obtains the frequency of words in that file.
def main():
  # Prompt user to enter file name and open it
  file_name = raw_input ("Enter file name: ")
  inFile = open (file_name, "r")

  # Create dictionary
  freq = {}
  
  # Read each line and process it
  for line in inFile:
    line = line.rstrip("\n")
    line = line.lower()
    
    # Remove punctuation marks
    line = line.replace(".", "")
    line = line.replace(",", "")
    line = line.replace(";", "")
    line = line.replace(":", "")
    line = line.replace("?", "")
    line = line.replace("!", "")
    line = line.replace("'", "")
    line = line.replace('"', "")

    # Split the line into individual words
    wordList = line.split()
    
    # Check each word and increment its count in the dictionary
    for word in wordList:
      if (word in freq):
        freq[word] = freq[word] + 1
      else:
        freq[word] = 1

  # Close file
  inFile.close()

  # Create reverse frequency dictionary
  rev_freq = {}

  # Process the frequency dictionary and add to rev_freq{}
  for word in freq:
    if (freq[word] in rev_freq):
      (rev_freq[freq[word]]).append(word)
    else:
      rev_freq[freq[word]] = [word]

  # Print out the frequency of words
  freqList = rev_freq.keys()
  freqList.sort()
  freqList.reverse()
  for count in freqList:
    wordList = rev_freq[count]
    wordList.sort()
    print count, wordList

main()