def interpolationLaplaceCalc():
train, y = preprocessing.train_split(
'text_files/outputs/preprocessing.txt')
del y
trainCount = Counter(train)
unigramCount = models.laplaceUnigram(train)
bigramCounts = Counter([(word, train[i + 1])
for i, word in enumerate(train[:-1])])
trigramCounts = Counter([(word, train[i + 1], train[i + 2])
for i, word in enumerate(train[:-2])])
first, last = endpoint.tools.stringParserNoCount(
request.args.get('firstword'), request.args.get('lastword'))
if first == "" and last == "":
probability = 0
else:
probability = pc.calculateProbabilityLaplaceInterpolation(
trainCount, unigramCount, bigramCounts, trigramCounts, first, last)
return render_template('%s.html' % "probabilityCalcInt",
value=str(probability),
type="laplaceInt",
flavour="laplace",
back="probCalcInt")
def interpolationUnkGen():
train, y = preprocessing.train_split(
'text_files/outputs/preprocessing.txt')
del y
unkUnigram = models.unkUnigram(train)
unkBigram = models.unkBigram(train)
unkTrigram = models.unkTrigram(train)
first, last, count = endpoint.tools.stringParser(
request.args.get('firstword'), request.args.get('lastword'),
request.args.get('count'))
sentence = tg.generateTextInterpolation(unkUnigram,
unkBigram,
unkTrigram,
first,
lastWord=last,
count=count)
fullSentence = ""
for word in sentence:
fullSentence += " " + word
return render_template('%s.html' % "textGenInt",
value=fullSentence,
type="unkInt",
flavour="unk",
back="testGenInt")
def laplaceTrigramGenerator():
train, y = preprocessing.train_split(
'text_files/outputs/preprocessing.txt')
del y
bigramCounts = Counter([(word, train[i + 1])
for i, word in enumerate(train[:-1])])
trigramCounts = Counter([(word, train[i + 1], train[i + 2])
for i, word in enumerate(train[:-2])])
first, last, count = endpoint.tools.stringParser(
request.args.get('firstword'), request.args.get('lastword'),
request.args.get('count'))
sentence = tg.generateTextFromLaplaceTrigram(train,
bigramCounts,
trigramCounts,
first,
lastWord=last,
count=count)
fullSentence = ""
for word in sentence:
fullSentence += " " + word
return render_template('%s.html' % "textGen",
value=fullSentence,
type="laplaceTrigram",
flavour="laplace",
back="testGen")
def vanillaTrigramGenerator():
train, y = preprocessing.train_split(
'text_files/outputs/preprocessing.txt')
del y
vanillaBigram = models.vanillaBigram(train)
vanillaTrigram = models.vanillaTrigram(train)
first, last, count = endpoint.tools.stringParser(
request.args.get('firstword'), request.args.get('lastword'),
request.args.get('count'))
sentence = tg.generateTextFromTrigram(vanillaBigram,
vanillaTrigram,
first,
lastWord=last,
count=count)
fullSentence = ""
for word in sentence:
fullSentence += " " + word
return render_template('%s.html' % "textGen",
value=fullSentence,
type="vanillaTrigram",
flavour="vanilla",
back="testGen")
def unkBigramCalc():
train, y = preprocessing.train_split(
'text_files/outputs/preprocessing.txt')
del y
unkBigram = models.unkBigram(train)
print(request.args.get('count'))
first, last = endpoint.tools.stringParserNoCount(
request.args.get('firstword'), request.args.get('lastword'))
if first == "" and last == "":
probability = 0
else:
probability = pc.calculateProbabilityFromBigram(unkBigram, first, last)
return render_template('%s.html' % "probabilityCalc",
value=str(probability),
type="unkBigram",
flavour="unk",
back="probCalc")
def interpolationVanillaCalc():
train, y = preprocessing.train_split(
'text_files/outputs/preprocessing.txt')
del y
vanillaUnigram = models.vanillaUnigram(train)
vanillaBigram = models.vanillaBigram(train)
vanillaTrigram = models.vanillaTrigram(train)
first, last = endpoint.tools.stringParserNoCount(
request.args.get('firstword'), request.args.get('lastword'))
if first == "" and last == "":
probability = 0
else:
probability = pc.calculateProbabilityInterpolation(
vanillaUnigram, vanillaBigram, vanillaTrigram, first, last)
return render_template('%s.html' % "probabilityCalcInt",
value=str(probability),
type="vanillaInt",
flavour="vanilla",
back="probCalcInt")
def interpolationlaplaceGen():
train, y = preprocessing.train_split(
'text_files/outputs/preprocessing.txt')
del y
vanillaBigram = models.vanillaBigram(train)
trainCount = Counter(train)
unigramCount = models.laplaceUnigram(train)
bigramCounts = Counter([(word, train[i + 1])
for i, word in enumerate(train[:-1])])
trigramCounts = Counter([(word, train[i + 1], train[i + 2])
for i, word in enumerate(train[:-2])])
first, last, count = endpoint.tools.stringParser(
request.args.get('firstword'), request.args.get('lastword'),
request.args.get('count'))
sentence = tg.generateTextLaplaceInterpolation(trainCount,
unigramCount,
bigramCounts,
vanillaBigram,
trigramCounts,
first,
lastWord=last,
count=count)
fullSentence = ""
for word in sentence:
fullSentence += " " + word
return render_template('%s.html' % "textGenInt",
value=fullSentence,
type="laplaceInt",
flavour="laplace",
back="testGenInt")
def main():
print("Started API")
x, y = preprocessing.train_split('text_files/outputs/preprocessing.txt')
