How many anagrams are there?

anagram.py

# Assume the average length m of each word << the length of the list of words, which we call n
# The following will run in O(n*mlgm) Now assuming that we're using words from 
# the english language, then the mlgm factor goes to around 6*C
# so the overall runtime should scale linearly with the size of the input length

def anagrams(file):
    dict = {}
    count = set()
    for line in open(file):
        sline = "".join(sorted(line))
        if sline in dict:
            count.add(sline) # change to line to find how many are part of the list of anagrams
        dict[sline] = 0
    return len(count)