def clean_urls_in_dictionary(self):
self.dictLock.acquire()
newdict = copy.deepcopy(self.DEFAULT_DICTIONARY)
for key in self.dictionary:
firsts = self.dictionary.get(key)[0]
for i in range(0, len(firsts)):
firsts[i] = (clean_url(firsts[i][0]), firsts[i][1])
seconds = self.dictionary.get(key)[1]
for i in range(0, len(seconds)):
seconds[i] = (clean_url(seconds[i][0]), seconds[i][1])
newkey = clean_url(key.split()[0])
if len(key.split()) > 1:
newkey = newkey + ' ' + clean_url(key.split()[1])
newdict[newkey] = (firsts, seconds)
self.dictionary = newdict
self.dictLock.release()
def clean_urls_in_dictionary(self):
self.dictLock.acquire()
newdict = copy.deepcopy(self.DEFAULT_DICTIONARY)
for key in self.dictionary:
firsts = self.dictionary.get(key)[0]
for i in range(0, len(firsts)):
firsts[i] = (clean_url(firsts[i][0]), firsts[i][1])
seconds = self.dictionary.get(key)[1]
for i in range(0, len(seconds)):
seconds[i] = (clean_url(seconds[i][0]), seconds[i][1])
newkey = clean_url(key.split()[0])
if len(key.split()) > 1:
newkey = newkey + ' ' + clean_url(key.split()[1])
newdict[newkey] = (firsts, seconds)
self.dictionary = newdict
self.dictLock.release()
def generate_chain(self, message):
"""Generates a Markov chain from a message"""
self.dictLock.acquire()
words = message.split()
words.append(self.STOPWORD)
words.insert(0, self.STOPWORD)
# find URLs, neaten them up
for i in range(0, len(words)):
words[i] = clean_url(words[i])
if '<{}>'.format(self.users[self.BOT_ID]) in words[1]:
del words[1]
if len(words) < 2:
return ''
# try to guess which word is the most important
subject = self.STOPWORD
confidence = 0
for word in words:
if self.wordcounts.has_key(word):
tfidf = tf_idf(word, words, self.wordcounts, self.sentences_ever)
if tfidf > confidence:
confidence = tfidf
subject = word
# pick a word pair we've seen used with that word before as a seed
pairs = []
for wordpair in self.paircounts:
temp = wordpair.split()
if (temp[0] == subject) or ((len(temp) > 1) and (temp[1] == subject)):
pairs.append((wordpair, self.paircounts.get(wordpair)))
seed = choose_word_from_list(pairs)
chain = ''
# forwards
wordpair = seed
if self.dictionary.has_key(wordpair):
chain = wordpair
#print wordpair
while (wordpair.split()[1] != self.STOPWORD) and (self.dictionary.has_key(wordpair)):
wordpair = wordpair.split()[1] + u' ' + \
choose_word_from_list(self.dictionary.get(wordpair)[1])
#print wordpair
chain = chain + u' ' + wordpair.split()[1]
# backwards
wordpair = seed
if self.dictionary.has_key(wordpair) and wordpair.split()[0] != self.STOPWORD:
wordpair = choose_word_from_list(
self.dictionary.get(wordpair)[0]) + \
u' ' + wordpair.split()[0]
# so we don't have the seed twice
while (wordpair.split()[0] != self.STOPWORD) and (self.dictionary.has_key(wordpair)):
#print wordpair
chain = wordpair.split()[0] + u' ' + chain
wordpair = choose_word_from_list(
self.dictionary.get(wordpair)[0]) + \
u' ' + wordpair.split()[0]
self.dictLock.release()
return chain.replace(self.STOPWORD, u'')
def interpret_message(self, message):
"""Interprets a message"""
self.dictLock.acquire()
words = message.split()
words.append(self.STOPWORD)
words.insert(0, self.STOPWORD)
self.sentences_ever = self.sentences_ever + 1
# find URLs, neaten them up
for i in range(0, len(words)):
words[i] = clean_url(words[i])
for word in words:
if not (self.wordcounts.has_key(word)):
self.wordcounts[word] = 0
self.wordcounts[word] = self.wordcounts.get(word) + 1
index = 0
word = words[index]
# cannot be out of range; at least (stop, stop, word, stop, stop)
wordpair = words[index] + u' ' + words[index + 1]
while True:
try:
next = words[index + 2]
nextpair = words[index + 1] + u' ' + words[index + 2]
except IndexError:
# this means we got to the end of the sentence
break
if not (self.paircounts.has_key(wordpair)):
self.paircounts[wordpair] = 0
self.paircounts[wordpair] = self.paircounts.get(wordpair) + 1
# add 'next' as a word that comes after 'wordpair'
if self.dictionary.has_key(wordpair):
temp = self.dictionary.get(wordpair)[1]
wordindex = word_index_in_list(next, temp)
if wordindex == -1:
temp.append((next, 1))
else:
prevcount = temp[wordindex][1]
temp[wordindex] = (next, prevcount + 1)
else:
self.dictionary[wordpair] = ([], [(next, 1)])
# add 'word' as a word that comes before 'nextpair'
if self.dictionary.has_key(nextpair):
othertemp = self.dictionary.get(nextpair)[0]
wordindex = word_index_in_list(word, othertemp)
if wordindex == -1:
othertemp.append((word, 1))
else:
prevcount = othertemp[wordindex][1]
othertemp[wordindex] = (word, prevcount + 1)
else:
self.dictionary[nextpair] = ([(word, 1)], [])
index = index + 1
word = words[index]
wordpair = word + u' ' + words[index + 1]
#print self.dictionary
self.dictLock.release()
def do_commands(self, target, sender, message, sentByAdmin):
if sentByAdmin and ('!saveDict' in message):
try:
self.save_dictionary()
self.send_message(target, 'DICTIONARY SAVED SUCCESSFULLY (%s bytes)' % str(os.path.getsize('Markov_Dict.pkl')))
except IOError:
self.send_message(target, 'DICTIONARY COULD NOT BE SAVED')
return True
elif sentByAdmin and ('!loadDict' in message):
try:
self.load_dictionary()
self.send_message(target, 'DICTIONARY LOADED SUCCESSFULLY (%s bytes)' % str(os.path.getsize('Markov_Dict.pkl')))
except IOError:
self.send_message(target, 'DICTIONARY COULD NOT BE LOADED')
return True
elif sentByAdmin and ('!eraseDict' in message):
self.dictionary = {
self.STOPWORD : ([self.STOPWORD], [self.STOPWORD])
}
self.send_message(target, 'DICTIONARY ERASED (NOT SAVED YET)')
return True
elif sentByAdmin and ('!learn' in message):
self.toggle_learn()
print_message = 'I AM {} LEARNING'
self.send_message(target,
print_message.format('NOW' if self.isLearning else 'NO LONGER'))
return True
elif sentByAdmin and ('!cleanURL' in message):
self.clean_urls_in_dictionary()
self.send_message(target, 'LINKS IN DICTIONARY HAVE BEEN CLEANED')
return True
elif '!search' in message:
try:
message = message.lower()
searchterms = message.split()[1:]
for i in range(0, len(searchterms)):
searchterms[i] = clean_url(searchterms[i])
if len(searchterms) == 1:
phrases = []
for key in self.dictionary:
if searchterms[0] == key.split()[0] or \
(len(key.split()) > 1 and \
searchterms[0] == key.split()[1]):
phrases.append(key)
self.send_message(target, u'"%s" in pairs: %s' % (searchterms[0], str(phrases)))
else:
key = searchterms[0] + u' ' + searchterms[1]
if self.dictionary.has_key(key):
self.send_message(target, u'"%s": %s' % (key, str(self.dictionary.get(key))))
else:
self.send_message(target, u'"%s" not found in dictionary' % key)
except IndexError:
self.send_message(target, u'MALFORMED COMMAND')
return True
elif '!talkback' in message:
try:
self.talkBackFreq = float(message.split()[1])
self.send_message(target, ('RESPONDING PROBABILITY SET TO %3f' % self.talkBackFreq))
except (IndexError, TypeError, ValueError):
self.send_message(target, 'MALFORMED COMMAND')
return True
elif sentByAdmin and ('!quit' in message):
self.quit()
return True
elif '!avatar' in message:
self.send_message(target, 'SOURCE OF MY CURRENT AVATAR: %s' % self.AVATARSOURCE)
return True
elif ('!nowplaying' in message):
songname, songartist = self.generate_song()
self.send_message(target, u'Now Playing: "%s", by %s' % (string.capwords(songname), string.capwords(songartist)))
return True
return False # did not find a command
def generate_chain(self, message):
"""Generates a Markov chain from a message"""
self.dictLock.acquire()
words = message.split()
words.append(self.STOPWORD)
words.insert(0, self.STOPWORD)
# find URLs, neaten them up
for i in range(0, len(words)):
words[i] = clean_url(words[i])
if '<{}>'.format(self.users[self.BOT_ID]) in words[1]:
del words[1]
if len(words) < 2:
return ''
# try to guess which word is the most important
subject = self.STOPWORD
confidence = 0
for word in words:
if self.wordcounts.has_key(word):
tfidf = tf_idf(word, words, self.wordcounts,
self.sentences_ever)
if tfidf > confidence:
confidence = tfidf
subject = word
# pick a word pair we've seen used with that word before as a seed
pairs = []
for wordpair in self.paircounts:
temp = wordpair.split()
if (temp[0] == subject) or ((len(temp) > 1) and
(temp[1] == subject)):
pairs.append((wordpair, self.paircounts.get(wordpair)))
seed = choose_word_from_list(pairs)
chain = ''
# forwards
wordpair = seed
if self.dictionary.has_key(wordpair):
chain = wordpair
#print wordpair
while (wordpair.split()[1] !=
self.STOPWORD) and (self.dictionary.has_key(wordpair)):
wordpair = wordpair.split()[1] + u' ' + \
choose_word_from_list(self.dictionary.get(wordpair)[1])
#print wordpair
chain = chain + u' ' + wordpair.split()[1]
# backwards
wordpair = seed
if self.dictionary.has_key(
wordpair) and wordpair.split()[0] != self.STOPWORD:
wordpair = choose_word_from_list(
self.dictionary.get(wordpair)[0]) + \
u' ' + wordpair.split()[0]
# so we don't have the seed twice
while (wordpair.split()[0] !=
self.STOPWORD) and (self.dictionary.has_key(wordpair)):
#print wordpair
chain = wordpair.split()[0] + u' ' + chain
wordpair = choose_word_from_list(
self.dictionary.get(wordpair)[0]) + \
u' ' + wordpair.split()[0]
self.dictLock.release()
return chain.replace(self.STOPWORD, u'')
def interpret_message(self, message):
"""Interprets a message"""
self.dictLock.acquire()
words = message.split()
words.append(self.STOPWORD)
words.insert(0, self.STOPWORD)
self.sentences_ever = self.sentences_ever + 1
# find URLs, neaten them up
for i in range(0, len(words)):
words[i] = clean_url(words[i])
for word in words:
if not (self.wordcounts.has_key(word)):
self.wordcounts[word] = 0
self.wordcounts[word] = self.wordcounts.get(word) + 1
index = 0
word = words[index]
# cannot be out of range; at least (stop, stop, word, stop, stop)
wordpair = words[index] + u' ' + words[index + 1]
while True:
try:
next = words[index + 2]
nextpair = words[index + 1] + u' ' + words[index + 2]
except IndexError:
# this means we got to the end of the sentence
break
if not (self.paircounts.has_key(wordpair)):
self.paircounts[wordpair] = 0
self.paircounts[wordpair] = self.paircounts.get(wordpair) + 1
# add 'next' as a word that comes after 'wordpair'
if self.dictionary.has_key(wordpair):
temp = self.dictionary.get(wordpair)[1]
wordindex = word_index_in_list(next, temp)
if wordindex == -1:
temp.append((next, 1))
else:
prevcount = temp[wordindex][1]
temp[wordindex] = (next, prevcount + 1)
else:
self.dictionary[wordpair] = ([], [(next, 1)])
# add 'word' as a word that comes before 'nextpair'
if self.dictionary.has_key(nextpair):
othertemp = self.dictionary.get(nextpair)[0]
wordindex = word_index_in_list(word, othertemp)
if wordindex == -1:
othertemp.append((word, 1))
else:
prevcount = othertemp[wordindex][1]
othertemp[wordindex] = (word, prevcount + 1)
else:
self.dictionary[nextpair] = ([(word, 1)], [])
index = index + 1
word = words[index]
wordpair = word + u' ' + words[index + 1]
#print self.dictionary
self.dictLock.release()
def do_commands(self, target, sender, message, sentByAdmin):
if sentByAdmin and ('!saveDict' in message):
try:
self.save_dictionary()
self.send_message(
target, 'DICTIONARY SAVED SUCCESSFULLY (%s bytes)' %
str(os.path.getsize('Markov_Dict.pkl')))
except IOError:
self.send_message(target, 'DICTIONARY COULD NOT BE SAVED')
return True
elif sentByAdmin and ('!loadDict' in message):
try:
self.load_dictionary()
self.send_message(
target, 'DICTIONARY LOADED SUCCESSFULLY (%s bytes)' %
str(os.path.getsize('Markov_Dict.pkl')))
except IOError:
self.send_message(target, 'DICTIONARY COULD NOT BE LOADED')
return True
elif sentByAdmin and ('!eraseDict' in message):
self.dictionary = {
self.STOPWORD: ([self.STOPWORD], [self.STOPWORD])
}
self.send_message(target, 'DICTIONARY ERASED (NOT SAVED YET)')
return True
elif sentByAdmin and ('!learn' in message):
self.toggle_learn()
print_message = 'I AM {} LEARNING'
self.send_message(
target,
print_message.format(
'NOW' if self.isLearning else 'NO LONGER'))
return True
elif sentByAdmin and ('!cleanURL' in message):
self.clean_urls_in_dictionary()
self.send_message(target, 'LINKS IN DICTIONARY HAVE BEEN CLEANED')
return True
elif '!search' in message:
try:
message = message.lower()
searchterms = message.split()[1:]
for i in range(0, len(searchterms)):
searchterms[i] = clean_url(searchterms[i])
if len(searchterms) == 1:
phrases = []
for key in self.dictionary:
if searchterms[0] == key.split()[0] or \
(len(key.split()) > 1 and \
searchterms[0] == key.split()[1]):
phrases.append(key)
self.send_message(
target,
u'"%s" in pairs: %s' % (searchterms[0], str(phrases)))
else:
key = searchterms[0] + u' ' + searchterms[1]
if self.dictionary.has_key(key):
self.send_message(
target,
u'"%s": %s' % (key, str(self.dictionary.get(key))))
else:
self.send_message(
target, u'"%s" not found in dictionary' % key)
except IndexError:
self.send_message(target, u'MALFORMED COMMAND')
return True
elif '!talkback' in message:
try:
self.talkBackFreq = float(message.split()[1])
self.send_message(
target,
('RESPONDING PROBABILITY SET TO %3f' % self.talkBackFreq))
except (IndexError, TypeError, ValueError):
self.send_message(target, 'MALFORMED COMMAND')
return True
elif sentByAdmin and ('!quit' in message):
self.quit()
return True
elif '!avatar' in message:
self.send_message(
target, 'SOURCE OF MY CURRENT AVATAR: %s' % self.AVATARSOURCE)
return True
elif ('!nowplaying' in message):
songname, songartist = self.generate_song()
self.send_message(
target, u'Now Playing: "%s", by %s' %
(string.capwords(songname), string.capwords(songartist)))
return True
return False # did not find a command
