def check_candidate_equality(dir_path1, dir_path2, cutoff_list, lambda_list):
dir1_dict = dict()
dir2_dict = dict()
fileset1 = set(listdir(dir_path1))
fileset2 = set(listdir(dir_path2))
for file_name in fileset1:
file_path = "%s/%s" % (dir_path1, file_name)
dir1_dict[file_name] = read_candidate(file_path, cutoff_list[0], cutoff_list[-1])
for file_name in fileset2:
file_path = "%s/%s" % (dir_path2, file_name)
dir2_dict[file_name] = read_candidate(file_path, cutoff_list[0], cutoff_list[-1])
diff = fileset1 - fileset2
intersection = fileset1.intersection(fileset2)
if diff:
print "These files do not exist in %s" % (dir_path2)
print diff
else:
for file_name in intersection:
for cutoff in cutoff_list:
for lamb in lambda_list:
list_check_equality(
file_name, dir1_dict[file_name][cutoff][lamb], dir2_dict[file_name][cutoff][lamb]
)
def run_evaluation(self, ref):
gold_standard = ref
print ref.keys()
eval_score = self.init_eval_score()
for k1 in self.k1_list:
for b in self.b_list:
candidate_dir_path = '../graph_old/%s_%s' % (str(k1),str(b))
file_set = set(listdir(candidate_dir_path))
for file_name in file_set:
candidates_dict = read_candidate(candidate_dir_path + '/' + file_name, self.cutoff_list[0], self.cutoff_list[-1])
file_name = '.'.join(re.split('\.', file_name)[:-1])
if file_name not in gold_standard.keys():
print 'ref file<%s> does not exist in the given gold_standard data' % (file_name)
self.excluded_files.add(file_name)
continue
gs_list = gold_standard[file_name]
for cutoff in self.cutoff_list:
for lamb in self.lamb_list[1:]:
if not candidates_dict[cutoff].get(lamb, 0): continue
candidates = candidates_dict[cutoff][lamb]
eval_score[cutoff][lamb][k1][b][file_name] = (
evaluate(self.stem_list(candidates), self.stem_list(gs_list)))
return eval_score