# reads in the text file and prints a list of words in the file
# repeated words are only printed once, and the list is sorted
# ECS 10, May 18, 2009
# Matt Bishop
import string

# these are what we care about
# a "word" is a maximal string of letters and digits
# whitespace separates words
# everything else is treated like whitespace
lettdig = string.letters + string.digits + string.whitespace

# returns the line with all non-lettdigs replaced by white space
# parameters: line, the line to be transformed
# returns: res, the transformed line
def letteronly(line):
        # initialize string to be returned
        res = ""
        # go through the line character by character
        for i in line:
                # if a lettdig, append it to result
                # if anything else, append a whitespace
                if i in lettdig:
                        res += i
                else:
                        res += " "
        # return the transformed string
        return res

# adds words to a list of words, without duplication
# note the adding is done in place
# parameters: wordlist, list of words being build
#             wl, list of words to add
def addwords(wordlist, wl):
    for i in wl:
        wordlist[i] = wordlist.get(i, 0) + 1

# comparison function to sort list of items
# parameters: (k1, v1), first item
#             (k2, v2), second item
# returns: -1 if v1 > v2, 0 if v1 == v2, 1 if v1 < v2
def cmpfun((k1, v1), (k2, v2)):
        if v1 > v2:
                return -1
        elif v1 == v2:
                return 0
        else:
                return 1

# this puts it all together
def main():
        # get the file name and open it
        try:
                fname = raw_input("File name: ")
                infile = open(fname, "r")
        except IOError:
                print "Could not open", fname
                return
        except EOFError:
                return
        # now initialize the dictionary
        wordlist = {}
        # loop through the file, one line at a time
        for l in infile:
                # break the line up into words (as defined above)
                bareline = letteronly(l)
                wordsinline = string.split(bareline)
                # add each new word to the word list
                addwords(wordlist, wordsinline)

        # all done! sort the word list
        # and print the words, one per line
        # this sorts them alphabetically
        itemlist = wordlist.items()
        itemlist.sort(cmpfun)
        for i in itemlist:
                print "%3d  %s" % (i[1], i[0])

main()