def generate_inputs_from_constraints(constraint_dict,
min_size,
max_string_size=100):
#sample a size from min to max
size = random.randint(min_size, max_string_size)
indices = set(range(size))
slist = random.choices(printable[:-4], k=size)
# schematically:
#print("min_size", min_size)
#print("size", size)
for item in constraint_dict:
#print("ITEM", item)
#print("sliststr:", ''.join(slist))
num_to_insert = max(
0, constraint_dict[item] -
len(re.findall(re.escape(item), ''.join(slist))))
#print("num_to_insert", num_to_insert)
if len(indices) < num_to_insert: return None
indices_to_insert = set(random.sample(indices, k=num_to_insert))
# do something here
#print("PREG INPUT",item if item not in preg_dict else preg_dict[item])
for i in indices_to_insert:
slist[i] = pre.create(
item).sample() if item not in preg_dict else pre.create(
preg_dict[item]).sample()
indices = indices - indices_to_insert
#may be too big but whatever
string = ''.join(slist)
if len(string) > max_string_size:
return string[:max_string_size] # may break but whatever
return string
def match_col(dataset, rstring):
r = pregex.create(rstring)
matches = []
for i, col in enumerate(dataset):
score = sum([r.match(example) for example in col])
if score != float('-inf'):
matches.append(i)
return matches
def regex_plus_bound(X):
from pregex import pregex
c = Counter(X)
regexes = [
pregex.create(".+"),
pregex.create("\d+"),
pregex.create("\w+"),
pregex.create("\s+"),
pregex.create("\\u+"),
pregex.create("\l+")]
regex_scores = []
for r in regexes:
regex_scores.append(sum(c[x] * r.match(x) for x in c)/float(sum([len(x) for x in X])) )
return max(regex_scores)