class Essay:
def __init__(self, filename, text, real_score = 0):
self.text = text
self.text_list = self.text.replace('\n'," ").split()
self.tags = nl.pos_tag(self.text_list)
self.raw_score = Score()
self.normalized_score = Score()
self.real_score = real_score
self.filename = filename
self.grader_score = 0.0
self.classification = 1
self.classification_penalty = 0
# Calculates the weighted score for an essay
def calculate_score(self):
s = self.normalized_score
self.grader_score = s.spelling + s.sbj_vrb + s.vrb_tense + 2*s.sent_form + 2*s.coherence + 3*s.topic + 2*s.length
return None
# Selects the class of the essay based on the learned cutoffs
def classify(self):
if self.grader_score > NormalizedCutoffs.total[0]:
self.classification += 1
if self.grader_score > NormalizedCutoffs.total[1]:
self.classification += 1
return None
# Replaces VBZ tags with VBZis or VBZhas
def disambiguate_vbz(self):
tags = map(lambda x : (x[0], 'VBZis') if x[1] == 'VBZ' and x[0] == 'is' else x, self.tags)
self.tags_vbz = map(lambda x : (x[0], 'VBZhas') if x[1] == 'VBZ' and x[0] == 'has' else x, tags)
return None
# Returns a string formatted for the competition
def output(self):
t = self.classification
if t == 1:
grade = "low"
elif t == 2:
grade = "medium"
else:
grade = "high"
string = self.filename + "\t" + str(self.normalized_score.spelling) + "\t" + str(self.normalized_score.sbj_vrb) + "\t" + str(self.normalized_score.vrb_tense) + "\t" + str(self.normalized_score.sent_form) + "\t" + str(self.normalized_score.anaphora) + "\t" + str(self.normalized_score.topic) + "\t" + str(self.normalized_score.length) + "\t" + grade + "\n"
return string
# Builds a parse tree using Pattern.en
def parse(self):
self.parsetree = parsetree(self.text, relations=True)
return None
# Sends the essay down the pipeline to calculate it's raw scores for each category
def pipeline(self):
print "Processing " + self.filename
self.raw_score.spelling = spelling.mistakes(self)
self.raw_score.sbj_vrb = sva.mistakes(self)
self.raw_score.length = sentence.length(self)
self.raw_score.vrb_tense = verb.mistakes(self)
self.raw_score.sent_form = sentenceFormation.mistakes(self)
self.raw_score.anaphora = centering.compute_ambiguity(self)
self.raw_score.topic = topicCoherence.score(self)
# Now predict
self.predict()
return None
# Takes an essay and scores it according to the trained cutoff points
def predict(self):
for cat in categories.ALL:
raw = getattr(self.raw_score, cat)
score = 1
# Score the essay according to its cutoff level
if cat == "length" or cat == "topic":
for i, cutoff in enumerate(getattr(NormalizedCutoffs, cat)):
if raw > cutoff:
score += 1
else:
for i, cutoff in enumerate(getattr(NormalizedCutoffs, cat)):
if raw < cutoff:
score += 1
# Store the normalized score
setattr(self.normalized_score, cat, score)
# Calculate the essay's total score and classify
self.calculate_score()
self.classify()
# Calculate mis-classification penalty
if self.real_score != 0 and self.real_score != self.classification:
self.classification_penalty = abs(self.real_score - self.classification)
return None
# Prints a summary of the essay
def summary(self):
print self.filename
print "Real Score: " + str(self.real_score)
print "Classification: " + str(self.classification)
print "Grader Score: " + str(self.grader_score)
self.normalized_score._print()
return None
class Essay:
def __init__(self, filename, text, real_score=0):
self.text = text
self.text_list = self.text.replace('\n', " ").split()
self.tags = nl.pos_tag(self.text_list)
self.raw_score = Score()
self.normalized_score = Score()
self.real_score = real_score
self.filename = filename
self.grader_score = 0.0
self.classification = 1
self.classification_penalty = 0
# Calculates the weighted score for an essay
def calculate_score(self):
s = self.normalized_score
self.grader_score = s.spelling + s.sbj_vrb + s.vrb_tense + 2 * s.sent_form + 2 * s.coherence + 3 * s.topic + 2 * s.length
return None
# Selects the class of the essay based on the learned cutoffs
def classify(self):
if self.grader_score > NormalizedCutoffs.total[0]:
self.classification += 1
if self.grader_score > NormalizedCutoffs.total[1]:
self.classification += 1
return None
# Replaces VBZ tags with VBZis or VBZhas
def disambiguate_vbz(self):
tags = map(
lambda x: (x[0], 'VBZis')
if x[1] == 'VBZ' and x[0] == 'is' else x, self.tags)
self.tags_vbz = map(
lambda x: (x[0], 'VBZhas')
if x[1] == 'VBZ' and x[0] == 'has' else x, tags)
return None
# Returns a string formatted for the competition
def output(self):
t = self.classification
if t == 1:
grade = "low"
elif t == 2:
grade = "medium"
else:
grade = "high"
string = self.filename + "\t" + str(
self.normalized_score.spelling
) + "\t" + str(self.normalized_score.sbj_vrb) + "\t" + str(
self.normalized_score.vrb_tense) + "\t" + str(
self.normalized_score.sent_form) + "\t" + str(
self.normalized_score.anaphora) + "\t" + str(
self.normalized_score.topic) + "\t" + str(
self.normalized_score.length) + "\t" + grade + "\n"
return string
# Builds a parse tree using Pattern.en
def parse(self):
self.parsetree = parsetree(self.text, relations=True)
return None
# Sends the essay down the pipeline to calculate it's raw scores for each category
def pipeline(self):
print "Processing " + self.filename
self.raw_score.spelling = spelling.mistakes(self)
self.raw_score.sbj_vrb = sva.mistakes(self)
self.raw_score.length = sentence.length(self)
self.raw_score.vrb_tense = verb.mistakes(self)
self.raw_score.sent_form = sentenceFormation.mistakes(self)
self.raw_score.anaphora = centering.compute_ambiguity(self)
self.raw_score.topic = topicCoherence.score(self)
# Now predict
self.predict()
return None
# Takes an essay and scores it according to the trained cutoff points
def predict(self):
for cat in categories.ALL:
raw = getattr(self.raw_score, cat)
score = 1
# Score the essay according to its cutoff level
if cat == "length" or cat == "topic":
for i, cutoff in enumerate(getattr(NormalizedCutoffs, cat)):
if raw > cutoff:
score += 1
else:
for i, cutoff in enumerate(getattr(NormalizedCutoffs, cat)):
if raw < cutoff:
score += 1
# Store the normalized score
setattr(self.normalized_score, cat, score)
# Calculate the essay's total score and classify
self.calculate_score()
self.classify()
# Calculate mis-classification penalty
if self.real_score != 0 and self.real_score != self.classification:
self.classification_penalty = abs(self.real_score -
self.classification)
return None
# Prints a summary of the essay
def summary(self):
print self.filename
print "Real Score: " + str(self.real_score)
print "Classification: " + str(self.classification)
print "Grader Score: " + str(self.grader_score)
self.normalized_score._print()
return None
