How many 3-letter prefixes are commonly used in English?

prefixes.py

'''
PROBLEM: 
  How many 3-letter prefixes are commonly used in English?

MOTIVATION:
  The Lumosity word game constantly tests my vocabulary and ability to 
  remember simple, common words.  I would like to improve my performance.

SOLUTION:
  Count the n-letter prefixes used in a dictionary.

USER PARAMS:
  - PREFIX_N - length of prefix in chars
  - PERC_THRESHOLD - output the top x percent of prefixes (range: 0.0-1.0)
  - WORDS_FILE - a list of dictionary words
  - OUTPUT_FILE - output to top prefixes to the given file path; or if set to 
      None, output to stdout
'''
PREFIX_N = 3
PERC_THRESHOLD = .10
WORDS_FILE = '/usr/share/dict/web2'
OUTPUT_FILE = 'output.txt'

# 0. load words file
f = open(WORDS_FILE)
lines = f.read()
f.close()
words = lines.split('\n')

# 1. gather common prefixes
prefixes = {}
for i in words: 
	if len(i) > PREFIX_N:
		pre = i[:PREFIX_N].lower()
		if prefixes.has_key(pre):
			prefixes[pre] += 1
		else:
			prefixes[pre] = 1

# 2. sort the prefixes by decreasing frequency
prefixes_by_freq = [(prefixes[i], i) for i in prefixes]
prefixes_by_freq.sort(reverse=True)

# 3. output the top x% of words using threshold
# for my words file:
# THRESHOLD  # PREFIXES  % OF WORDS 
#         3         114       37.14
#         5         190       47.19
#        10         381       63.29
#        25         953       85.27
#        50        1906       96.83
#        75        2859       99.43
total_counts = float(sum([i[0] for i in prefixes_by_freq]))
lmt = int( len(prefixes_by_freq) * PERC_THRESHOLD )
output_lines = []
for i in prefixes_by_freq[:lmt]:
	freq = str(i[0])
	pre = i[1]
	perc = str(i[0] / total_counts * 100)
	output_lines.append( freq + "\t" + pre + "\t" + perc )
if OUTPUT_FILE:
	f = open(OUTPUT_FILE, 'wb')
	for i in output_lines:
		f.write(i + "\n")
	f.close()
else:
	for i in output_lines:
		print i