def parse_and_eval(string, values=[]):
expr, variables = parse_webwork(string)
try:
if expr:
if len(variables) == 0:
etree = eval_parsed(expr)
return etree, {}
if len(variables) != len(values):
return expr, variables # if the number of values does
# not match number of variables,
# return the description of the
# variables.
else:
# if number of values matches
# number of variables then
# substitute values for
# variables and call parser
# again on numerical expression.
subs_string = substitute(string, variables, values)
subs_expr, variables = parse_webwork(subs_string)
assert len(variables) == 0
if subs_expr:
etree = eval_parsed(subs_expr)
return etree, {}
else:
raise Exception
except:
log.error('Exception in parse_and_eval, string=%s, parsed=%s' %
(string, str(expr)))
traceback.print_exc(file=sys.stdout)
return None
def find_matches(answer, attempt):
logger.debug('find_matches recieved answer=' + str(answer) + ' attempt=' +
str(attempt))
attempt_parsed = parse_webwork(attempt)
logger.debug('calling eval_parsed on on' + str(attempt_parsed))
attempt_tree = eval_parsed(attempt_parsed)
logger.debug('calling flatten on' + str(attempt_tree))
attempt_list = flatten(attempt_tree, 't')
logger.debug('attempt list:\n' + str(attempt_list))
answer_parsed = parse_webwork(answer)
logger.debug('calling eval_parsed on on' + str(answer_parsed))
answer_tree = eval_parsed(answer_parsed)
logger.debug('calling flatten on' + str(answer_tree))
answer_list = flatten(answer_tree, 'c')
logger.debug('answer list\n' + str(answer_list))
combined_list = sorted(answer_list + attempt_list, key=lambda x: x[0])
logger.debug('combined list\n' + str(combined_list))
Hits = find_Hits(combined_list)
logger.debug('Hits:\n' + str(Hits))
final_matches = find_dominating_hits(Hits, answer, attempt)
return final_matches
def parse_and_eval(string,values=[]):
expr,variables = parse_webwork(string)
try:
if expr:
if len(variables)==0:
etree = eval_parsed(expr)
return etree,{}
if len(variables)!=len(values):
return expr,variables # if the number of values does
# not match number of variables,
# return the description of the
# variables.
else:
# if number of values matches
# number of variables then
# substitute values for
# variables and call parser
# again on numerical expression.
subs_string=substitute(string,variables,values)
subs_expr,variables=parse_webwork(subs_string)
assert len(variables)==0;
if subs_expr:
etree=eval_parsed(subs_expr)
return etree,{}
else:
raise Exception
except:
log.error('Exception in parse_and_eval, string=%s, parsed=%s'%(string,str(expr)))
traceback.print_exc(file=sys.stdout)
return None
def from_string(cls, string):
"""
Creates expression tree from a string.
Throws webwork_parser.WebworkParseException on failure
"""
data = webwork_parser.parse_webwork(string)
return cls(data)
def parse_and_collect_numbers(string):
try:
parse_tree = parse_webwork(string)
eval_tree = eval_parsed(parse_tree)
eval_numbers = Collect_numbers(eval_tree)
return set(eval_numbers.keys())
except:
return set()
def parse_and_collect_numbers(string):
try:
parse_tree = parse_webwork(string)
eval_tree = eval_parsed(parse_tree)
eval_numbers = Collect_numbers(eval_tree)
return set(eval_numbers.keys())
except:
return set()
def from_string(cls, string):
"""
Creates expression tree from a string.
Throws webwork_parser.WebworkParseException on failure
"""
data = webwork_parser.parse_webwork(string)
return cls(data)
def parse_and_eval(string, variables=None):
expr = parse_webwork(string)
if expr:
try:
etree = eval_parsed(expr, variables)
return expr, etree
except:
return (None, None)
else:
return (None, None)
def parse_and_eval(string, variables=None):
expr = parse_webwork(string)
if expr:
try:
etree = eval_parsed(expr, variables)
return expr, etree
except:
return (None, None)
else:
return (None, None)
def parse_and_collect_numbers(string):
try:
parse_tree, variables = parse_webwork(string)
if len(variables)==0:
eval_tree = eval_parsed(parse_tree)
eval_numbers = Collect_numbers(eval_tree)
return set(eval_numbers.keys())
else:
return Eval_with_Variables(string,variables)
except:
return set()
def parsed(string):
expr = parse_webwork(string)
if expr:
try:
etree = eval_parsed(expr)
nums = Collect_numbers(etree)
return etree, nums
except:
return (None, None)
else:
return (None, None)
def parsed(string):
expr = parse_webwork(string)
if expr:
try:
etree = eval_parsed(expr)
nums = Collect_numbers(etree)
return etree, nums
except:
return (None, None)
else:
return (None, None)
def parse_and_collect_numbers(string):
try:
parse_tree, variables = parse_webwork(string)
if len(variables) == 0:
eval_tree = eval_parsed(parse_tree)
eval_numbers = Collect_numbers(eval_tree)
return set(eval_numbers.keys())
else:
return Eval_with_Variables(string, variables)
except:
return set()
def parse_eval(string):
""" Given an expression, return it's parse tree and it's evaluation tree """
expr = parse_webwork(string)
if expr:
try:
etree = eval_parsed(expr)
return expr, etree
except:
return (None, None)
else:
return (None, None)
def parse_eval(string):
""" Given an expression, return it's parse tree and it's evaluation tree """
expr = parse_webwork(string)
if expr:
try:
etree = eval_parsed(expr)
return expr, etree
except:
return (None, None)
else:
return (None, None)
def parse_pg_file(pg_file):
answer_re = re.compile(".*\[_+\]{(.*)}")
compute_re = re.compile("Compute\(\"(.*)\"\)")
with open(os.path.expanduser(pg_file), 'r') as f:
for line in f:
m = answer_re.match(line)
if m:
print line
print m.group(1)
mc = compute_re.match(m.group(1))
if mc:
print mc.group(1)
parsed = parse_webwork(mc.group(1))
print parsed
def parse_pg_file(pg_file):
answer_re = re.compile(".*\[_+\]{(.*)}")
compute_re = re.compile("Compute\(\"(.*)\"\)")
with open(os.path.expanduser(pg_file), 'r') as f:
for line in f:
m = answer_re.match(line)
if m:
print line
print m.group(1)
mc = compute_re.match(m.group(1))
if mc:
print mc.group(1)
parsed = parse_webwork(mc.group(1))
print parsed
def post(self):
'''
Parses an expression and returns the parsed data structure.
Sample arguments:
expression="1+1"
Response:
[['+', [0,2]], 1, 1]
'''
expression = self.get_argument('expression')
etree = parse_webwork(expression)
if etree:
self.write(json.dumps(etree))
else:
logger.error("Failed to parse expression '%s'", expression)
self.set_status(400)
self.write(json.dumps({'error': 'Error parsing expression %s' % expression}))
def post(self):
'''
Parses an expression and returns the parsed data structure.
Sample arguments:
expression="1+1"
Response:
[['+', [0,2]], 1, 1]
'''
expression = self.get_argument('expression')
etree = parse_webwork(expression)
if etree:
self.write(json.dumps(etree))
else:
logger.error("Failed to parse expression '%s'", expression)
self.set_status(400)
self.write(
json.dumps(
{'error': 'Error parsing expression %s' % expression}))
# e3=webwork_parser.parse_webwork('C(6*7*5,5^3)+9/(31+18)')
# print e3
# We want to cluster all incorrect attempts
# For each incorrect attempt:
# Parse answer expression into tree
# (Discard unparseable expressions or something)
# Stick parse tree into a list
trees = []
part = 0
for i, a in answers.iteritems():
if df.scores[i][part] == '0':
# Only consider incorrect answers
try:
expr = webwork_parser.parse_webwork(a[part])
n = ExpressionNode(expr)
trees.append(n)
except webwork_parser.WebworkParseException:
pass
print len(trees), 'incorrect expressions parsed'
"""
K Medoids Algorithm
1. Initialize: randomly select k of the n data points as the medoids
2. Associate each data point to the closest medoid.
3. For each medoid m
1. For each non-medoid data point o
def get(self):
'''
Parses all expressions for a given question
Sample arguments:
course='CSE103_Fall14',
set_id='Week1',
problem_id=1,
part_id=1
include_finished=0
Response:
...
'''
# Get correct answers
logger.debug('Starting get')
self.answer_exps = {}
course = self.get_argument('course')
set_id = self.get_argument('set_id')
problem_id = self.get_argument('problem_id')
part_id = int(self.get_argument('part_id'))
include_finished = (int(self.get_argument('include_finished', 1)) == 1)
source_file = conn.query('''
select source_file from {course}_problem
where
problem_id={problem_id} and
set_id="{set_id}";
'''.format(course=course, set_id=set_id, problem_id=problem_id))[0]['source_file']
pg_path = os.path.join(webwork_dir, 'courses', course, 'templates', source_file)
with open(pg_path, 'r') as fin:
pg_file = fin.read()
# name problem_id set_id user_id value
user_variables = conn.query('''SELECT * from {course}_user_variables
WHERE set_id="{set_id}" AND problem_id = {problem_id};
'''.format(course=course, set_id=set_id, problem_id=problem_id))
self.variables_df = pd.DataFrame(user_variables)
if len(self.variables_df) == 0:
logger.warn("No user variables saved for assignment %s, please run the save_answers script", set_id)
# raise tornado.web.HTTPError(500)
answer_re = re.compile('\[__+\]{(?:Compute\(")?(.+?)(?:"\))?}')
answer_boxes = answer_re.findall(pg_file)
self.part_answer = answer_boxes[part_id-1]
self.answer_ptree = parse_webwork(self.part_answer)
# Get attempts by part
if include_finished:
query = '''SELECT * from {course}_answers_by_part
WHERE set_id="{set_id}" AND problem_id = {problem_id}
AND part_id={part_id};
'''.format(course=course, set_id=set_id, problem_id=problem_id,
part_id=part_id)
else:
# This self join query idea comes from http://stackoverflow.com/a/4519302/90551
# You can do this more clearly with subqueries but it's super slow
query = '''SELECT abp.* FROM {course}_answers_by_part as abp
LEFT JOIN {course}_answers_by_part AS abp2
ON abp.problem_id = abp2.problem_id AND abp.set_id = abp2.set_id
AND abp.part_id = abp2.part_id AND abp.user_id = abp2.user_id AND abp2.score = '1'
WHERE abp.set_id='{set_id}' AND abp.problem_id={problem_id}
AND abp.part_id={part_id} AND abp2.user_id IS NULL;
'''.format(course=course, set_id=set_id, problem_id=problem_id,
part_id=part_id)
logger.debug('Before sending SQL')
answers = conn.query(query)
logger.debug('After sending SQL')
all_correct_terms = set()
correct_terms_map = defaultdict(lambda: defaultdict(list))
incorrect_terms_map = defaultdict(lambda: defaultdict(list))
# Parse and evaluate all answers
for a in answers:
# {'user_id': u'acpatel', 'timestamp': datetime.datetime(2014, 10, 12, 2, 10, 19), 'id': 284488L, 'score': u'1', 'answer_string': u'C(55,6)', 'part_id': 2L, 'problem_id': 10L, 'set_id': u'Week2', 'answer_id': 199326L}
user_id = a['user_id']
ans = self.answer_for_student(user_id)
if a['score'] != '1':
etree, nums = parsed(a['answer_string'])
if etree and nums:
correct_terms = self.correct_terms(nums, ans)
all_correct_terms |= set(correct_terms)
# logger.debug(set(correct_terms))
if a['user_id'] not in correct_terms_map[str(sorted(correct_terms))][a['answer_string']]:
correct_terms_map[str(sorted(correct_terms))][a['answer_string']].append(a['user_id'])
out = {}
out['correct'] = correct_terms_map
out['correct_terms'] = sorted(all_correct_terms)
out['incorrect'] = incorrect_terms_map
out['answer_ptree'] = str(self.answer_ptree)
self.write(json.dumps(out, default=serialize_datetime))
logger.debug('Finished')
# variables.
else:
# if number of values matches
# number of variables then
# substitute values for
# variables and call parser
# again on numerical expression.
subs_string = substitute(string, variables, values)
subs_expr, variables = parse_webwork(subs_string)
assert len(variables) == 0
if subs_expr:
etree = eval_parsed(subs_expr)
return etree, {}
else:
raise Exception
except:
log.error('Exception in parse_and_eval, string=%s, parsed=%s' %
(string, str(expr)))
traceback.print_exc(file=sys.stdout)
return None
if __name__ == "__main__":
string = sys.argv[1]
print 'input:::', string
tree, variables = parse_webwork(string)
print 'output:::', tree, variables
eval_tree = eval_parsed(tree)
print 'Eval_tree:::', eval_tree
def parse_and_collect_numbers(string):
try:
parse_tree = parse_webwork(string)
eval_tree = eval_parsed(parse_tree)
eval_numbers = Collect_numbers(eval_tree)
return set(eval_numbers.keys())
except:
return set()
def parse_and_eval(string, variables=None):
expr = parse_webwork(string)
if expr:
try:
etree = eval_parsed(expr, variables)
return expr, etree
except:
return (None, None)
else:
return (None, None)
if __name__ == "__main__":
string = sys.argv[1]
print 'input:::', string
tree = parse_webwork(string)
print 'output:::', tree
eval_tree = eval_parsed(tree)
print 'Eval_tree:::', eval_tree
# not match number of variables,
# return the description of the
# variables.
else:
# if number of values matches
# number of variables then
# substitute values for
# variables and call parser
# again on numerical expression.
subs_string=substitute(string,variables,values)
subs_expr,variables=parse_webwork(subs_string)
assert len(variables)==0;
if subs_expr:
etree=eval_parsed(subs_expr)
return etree,{}
else:
raise Exception
except:
log.error('Exception in parse_and_eval, string=%s, parsed=%s'%(string,str(expr)))
traceback.print_exc(file=sys.stdout)
return None
if __name__=="__main__":
string=sys.argv[1]
print 'input:::',string
tree,variables=parse_webwork(string)
print 'output:::',tree, variables
eval_tree=eval_parsed(tree)
print 'Eval_tree:::',eval_tree
def get(self):
'''
Parses all expressions for a given question
Sample arguments:
course='CSE103_Fall14',
set_id='Week1',
problem_id=1,
part_id=1
include_finished=0
Response:
...
'''
# Get correct answers
logger.debug('Starting get')
self.answer_exps = {}
course = self.get_argument('course')
set_id = self.get_argument('set_id')
problem_id = self.get_argument('problem_id')
part_id = int(self.get_argument('part_id'))
include_finished = (int(self.get_argument('include_finished', 1)) == 1)
source_file = conn.query('''
select source_file from {course}_problem
where
problem_id={problem_id} and
set_id="{set_id}";
'''.format(course=course, set_id=set_id,
problem_id=problem_id))[0]['source_file']
pg_path = os.path.join(webwork_dir, 'courses', course, 'templates',
source_file)
with open(pg_path, 'r') as fin:
pg_file = fin.read()
# name problem_id set_id user_id value
user_variables = conn.query('''SELECT * from {course}_user_variables
WHERE set_id="{set_id}" AND problem_id = {problem_id};
'''.format(course=course, set_id=set_id, problem_id=problem_id))
self.variables_df = pd.DataFrame(user_variables)
if len(self.variables_df) == 0:
logger.warn(
"No user variables saved for assignment %s, please run the save_answers script",
set_id)
# raise tornado.web.HTTPError(500)
answer_re = re.compile('\[__+\]{(?:Compute\(")?(.+?)(?:"\))?}')
answer_boxes = answer_re.findall(pg_file)
self.part_answer = answer_boxes[part_id - 1]
self.answer_ptree = parse_webwork(self.part_answer)
# Get attempts by part
if include_finished:
query = '''SELECT * from {course}_answers_by_part
WHERE set_id="{set_id}" AND problem_id = {problem_id}
AND part_id={part_id};
'''.format(course=course,
set_id=set_id,
problem_id=problem_id,
part_id=part_id)
else:
# This self join query idea comes from http://stackoverflow.com/a/4519302/90551
# You can do this more clearly with subqueries but it's super slow
query = '''SELECT abp.* FROM {course}_answers_by_part as abp
LEFT JOIN {course}_answers_by_part AS abp2
ON abp.problem_id = abp2.problem_id AND abp.set_id = abp2.set_id
AND abp.part_id = abp2.part_id AND abp.user_id = abp2.user_id AND abp2.score = '1'
WHERE abp.set_id='{set_id}' AND abp.problem_id={problem_id}
AND abp.part_id={part_id} AND abp2.user_id IS NULL;
'''.format(course=course,
set_id=set_id,
problem_id=problem_id,
part_id=part_id)
logger.debug('Before sending SQL')
answers = conn.query(query)
logger.debug('After sending SQL')
all_correct_terms = set()
correct_terms_map = defaultdict(lambda: defaultdict(list))
incorrect_terms_map = defaultdict(lambda: defaultdict(list))
# Parse and evaluate all answers
for a in answers:
# {'user_id': u'acpatel', 'timestamp': datetime.datetime(2014, 10, 12, 2, 10, 19), 'id': 284488L, 'score': u'1', 'answer_string': u'C(55,6)', 'part_id': 2L, 'problem_id': 10L, 'set_id': u'Week2', 'answer_id': 199326L}
user_id = a['user_id']
ans = self.answer_for_student(user_id)
if a['score'] != '1':
etree, nums = parsed(a['answer_string'])
if etree and nums:
correct_terms = self.correct_terms(nums, ans)
all_correct_terms |= set(correct_terms)
# logger.debug(set(correct_terms))
if a['user_id'] not in correct_terms_map[str(
sorted(correct_terms))][a['answer_string']]:
correct_terms_map[str(
sorted(correct_terms))][a['answer_string']].append(
a['user_id'])
out = {}
out['correct'] = correct_terms_map
out['correct_terms'] = sorted(all_correct_terms)
out['incorrect'] = incorrect_terms_map
out['answer_ptree'] = str(self.answer_ptree)
self.write(json.dumps(out, default=serialize_datetime))
logger.debug('Finished')
