def get_software_name_and_version_from_cpe(cpe):
# print(cpe)
parts = cpe.split(':')[2:]
# print(parts)
software, version = '', ''
num_found = False
idx = 0
for part in parts:
if not utils.contain_letter(part) and not utils.contain_number(part):
break
part = part.replace('_', ' ').replace('~', ' ')
if part[0].isdigit():
version += part + ' '
num_found = True
elif num_found:
version += part + ' '
else:
software += part + ' '
idx += 1
software = software.strip()
version = version.strip()
if version == '':
software, version = extract_windows_version(software)
if software == '':
software, version = corner_case(version)
return software, version
def judge_word_in_sentence_with_both_dots_and_numbers(sent):
words = nltk.word_tokenize(sent)
for word in words:
contains_number = utils.contain_number(word)
contains_dot = '.' in word
if contains_number and contains_dot:
return True
return False
def corner_case(version):
# '1024cms 1024 cms 1.4.2 beta'
version_split = version.split()
version_split.reverse()
idx = 0
for word in version_split:
if utils.contain_number(word):
version_split.reverse()
return ' '.join(version_split[:-idx - 1]), ' '.join(
version_split[-idx - 1:])
idx += 1
return '', version
def clean_redundant_words_and_reserve_range(before_clean_list):
# preserve words that are
# (1) in the cve word set or
# (2) contain numbers or
# (3) do not contain number and letter or
# (4) in range word set
# todo: enrich range word set
range_word_set = {'before', 'older', 'prior', 'up', 'to', 'through', 'and', 'earlier',
'upper', 'higher', 'lower', 'including', 'since', 'onwards'}
after_clean_set = set()
for version_str in before_clean_list:
clean_version_str = ''
version_str_word_list = version_str.split()
for word in version_str_word_list:
if utils.contain_number(word) or (not utils.contain_number(word) and not utils.contain_letter(
word)) or word in range_word_set:
clean_version_str += word + ' '
after_clean_set.add(clean_version_str.strip())
after_clean_set -= {''}
# if '' in after_clean_set:
# after_clean_set.remove('')
return list(after_clean_set)
def extract_pair_from_edb_title(raw_title):
title_dict = {}
if raw_title.find(' - ') != -1:
# software_in_title might contain version
software_in_title = raw_title.split(' - ')[0]
contains_number = utils.contain_number(software_in_title)
if contains_number:
content_line_version, content_line_software = get_pair_from_content_line_focus_official(
raw_title)
content_line_version = content_line_version[:content_line_version.
find(' - ')]
content_line_version, content_line_software = move_range_from_software_to_version(
content_line_version, content_line_software)
title_dict = {
encode_content(content_line_software):
encode_content(content_line_version)
}
return title_dict
def get_pair_from_content_line_focus_official(line):
line = line.lower()
con_word_list = line.split()
content_line_version, content_line_software = '', ''
keyword_software_loc = 0
# contains keyword software
word_idx = 0
mat1 = False
version_loc = -1
for word in con_word_list:
if word == '':
continue
if word in ['windows', 'office']:
version_loc = word_idx + 1
mat1 = True
break
word_idx += 1
if mat1:
content_line_version = get_right_part(con_word_list, version_loc)
content_line_software = get_left_part(con_word_list, version_loc)
else:
mat1 = False
# the n-th word is version
version_loc = 0
word_idx = 0
for word in con_word_list:
if word == '':
continue
mat1 = re.match(r'(v)?[\d]{1,2}((\.[\d]{1,2}){1,2}(\.x)?|\.x)',
word)
if mat1:
version_loc = word_idx
break
word_idx += 1
# contains 1.1.x format number
if version_loc != 0:
content_line_version = get_right_part(con_word_list, version_loc)
content_line_software = get_left_part(con_word_list, version_loc)
# content_line_software = con[:con.find(content_line_version)].strip()
# print(content_line_software, ' ||| ', content_line_version)
# print(con)
# print()
# find word that is a number
else:
version_loc = 0
word_idx = 0
for word in con_word_list:
if word == '':
continue
contains_number = utils.contain_number(word)
if contains_number and word.lower not in [
'x64', 'x86', 'x86_64'
]:
version_loc = word_idx
break
word_idx += 1
content_line_version = get_right_part(con_word_list, version_loc)
content_line_software = get_left_part(con_word_list, version_loc)
# print(content_line_software, ' ||| ', content_line_version)
# print(con)
# print()
if content_line_version in content_line_software:
print(
'ERROR if content_line_version in content_line_software:')
return content_line_version.lower(), remove_duplicate_word_from_software(
content_line_software).lower()
def remove_space_in_focus_version(version_str):
version_split = version_str.split()
if len(version_split) == 2:
if version_split[1][0] == '.' and utils.contain_number(version_split[1][1:]):
version_str = version_str.replace(' ', '')
return version_str
