def add2grams(wordDict=None, write=True):
if wordDict is None:
wordDict = files.readPickle("./populatedDict.p")
for text in getBooks():
name = text[text.rfind(os.sep) + 1:]
print "reading book: %s" % name
book = files.readFile(text)
print "generating the list of words"
words = generateWordList(book)
count = 0
for i in xrange(2, len(words) - 2):
if words[i] not in wordDict:
continue
wordDict[words[i]].addWordAfter(words[i + 2], second=True)
wordDict[words[i]].addWordAfter(words[i + 1], second=False)
wordDict[words[i]].addWordBefore(words[i - 2], second=True)
wordDict[words[i]].addWordBefore(words[i - 1], second=False)
if count % 10000:
print "%d words read" % count
if write:
files.writePickle("2gramDict.p", wordDict)
return wordDict
def populatePossesives(wordDict, force=False, ratio=10, debug=False):
print "populating possesives"
count = 0
size = 5000
for word in wordDict:
count += 1
if debug and count % size == 0:
print "%d words done" % count
if word.endswith("'s") and (force or wordDict[word].count == 0):
# this makes certain that word is a word that is a possive
# ie: "man's" is a possessive
# my current list of ngrams doesn't have many words in
# possesive form so I'll artificially give it a count
try:
wordDict[word].count = wordDict[word[:-2]].count / ratio
except:
if debug:
print "error at %s" % word
continue
return wordDict
if __name__ == "__main__":
wordDict = files.readPickle("./2gramDict.p")
print readWord(wordDict["argument's"])
lis = ["fun", "help", "what", "the", "equality", "parity", "argument",
"aardvark", "Alps", "elk"]
for word in lis:
print readWord(wordDict[word])
