def generate_personal_data(num_records):
person = Person('en')
payment = Payment()
dtype = np.dtype([
('id', np.unicode_, 16),
('full_name', np.unicode_, 32),
('credit_card_number', np.unicode_, 19),
('credit_card_expiration_date', np.unicode_, 7),
('gender', np.unicode_, 1),
])
np.random.seed()
rng = np.random.randint(0, 2, num_records)
records_npy = np.empty(num_records, dtype=dtype)
for idx in range(num_records):
gender = Gender.MALE if rng[idx] == 0 else Gender.FEMALE
records_npy['id'][idx] = str(idx).zfill(16)
records_npy['full_name'][idx] = person.full_name(gender=gender)
records_npy['credit_card_number'][idx] = payment.credit_card_number(
card_type=CardType.VISA)
records_npy['credit_card_expiration_date'][
idx] = payment.credit_card_expiration_date(maximum=21)
records_npy['gender'][idx] = 'M' if rng[idx] == 0 else 'F'
return records_npy
def generate_data(self):
np.random.seed(seed=self.seed)
gender = np.random.randint(2, size=self.num)
loan = np.random.randint(self.loan_min / self.loan_step,
self.loan_max / self.loan_step,
size=self.num) * self.loan_step
person = Person('en')
payment = Payment()
self.main_array = np.empty(shape=self.num, dtype=self.dtype)
for idx in range(self.num):
self.main_array['id'][idx] = str(idx).zfill(16)
self.main_array['full_name'][idx] = person.full_name(
gender=Gender.MALE if gender[idx] else Gender.FEMALE)
self.main_array['credit_card_number'][
idx] = payment.credit_card_number(card_type=CardType.VISA)
self.main_array['credit_card_expiration_date'][
idx] = payment.credit_card_expiration_date(maximum=22)
self.main_array['gender'][idx] = 'M' if gender[idx] else 'F'
self.main_array['loan'][idx] = loan[idx]
same_entries_indices = np.random.choice(np.arange(self.num),
self.number_same_name_entries,
replace=False)
self.main_array['full_name'][same_entries_indices] = self.main_array[
'full_name'][same_entries_indices[0]]
self.main_array['gender'][same_entries_indices] = self.main_array[
'gender'][same_entries_indices[0]] # makes more sense
class DataGenerator:
def __init__(self, seed=13, similar_people_count=5, max_repeat_count=12):
self._initialize_counts_seed(seed, similar_people_count,
max_repeat_count)
self.df_people = self._generate_people()
self.df_cards = self._generate_cards()
self.df_records = pd.merge(self.df_people,
self.df_cards,
left_on='id_person',
right_on='id_person')
def _initialize_counts_seed(self, seed, similar_people_count,
max_repeat_count):
self.seed = seed
self.similar_people_count = similar_people_count
self.max_repeat_count = max_repeat_count
self.records_count = self._compute_records_count()
np.random.seed(self.seed)
self.person = Person('en', seed=self.seed)
self.payment = Payment(seed=self.seed)
def _compute_records_count(self):
num_records = 0
for i in range(1, self.max_repeat_count + 1):
num_records += i * self.similar_people_count
return num_records
def get_records(self):
return self.df_records
def _generate_cards(self):
person_ids = iter(self._generate_ids())
int_loans = np.random.randint(1, 101, self.records_count)
float_loans = np.array(int_loans * 1000, np.float64)
loans = iter(float_loans)
description_c = (lambda: {
'id_person':
next(person_ids),
'credit_card_num':
self.payment.credit_card_number(card_type=CardType.VISA),
'credit_card_exp_date':
self.payment.credit_card_expiration_date(maximum=21, minimum=19),
'loan':
next(loans),
})
schema_card = Schema(schema=description_c)
cards = schema_card.create(iterations=self.records_count)
return pd.DataFrame(cards)
def _generate_ids(self):
person_ids = self.df_people['id_person'].tolist()
repeat_count = 0
ids_for_cards = []
for i in range(len(person_ids)):
if i % self.similar_people_count == 0 and repeat_count < self.max_repeat_count:
repeat_count = repeat_count + 1
ids_for_cards = ids_for_cards + [i] * repeat_count
return ids_for_cards
def _generate_people(self):
people_count = self.similar_people_count * self.max_repeat_count
ids = iter(range(people_count))
description_female = (lambda: {
'id_person': next(ids),
'full_name': self.person.full_name(Gender.FEMALE),
'gender': 'F',
})
description_male = (lambda: {
'id_person': next(ids),
'full_name': self.person.full_name(Gender.MALE),
'gender': 'M',
})
female_count = people_count // 2
male_count = people_count - female_count
schema_female = Schema(schema=description_female)
females = schema_female.create(iterations=female_count)
schema_male = Schema(schema=description_male)
males = schema_male.create(iterations=male_count)
return pd.DataFrame(females + males)
('credit_card_number', np.unicode_, 32),
('credit_card_expiration_date', np.unicode_, 8),
('gender', np.unicode_, 1),
])
num = 10**2
rng = np.random.randint(0, 2, num)
person = Person('en')
payment = Payment()
array = np.empty(num, dtype=dtype)
for idx in range(num):
gender = Gender.MALE if rng[idx] == 0 else Gender.FEMALE
array['id'][idx] = str(idx).zfill(16)
array['full_name'][idx] = person.full_name(gender=gender)
array['credit_card_number'][idx] = payment.credit_card_number(card_type=CardType.VISA)
array['credit_card_expiration_date'][idx] = payment.credit_card_expiration_date(maximum=21)
array['gender'][idx] = 'M' if rng[idx] == 0 else 'F'
print(array)
print(array['full_name'])
ar_id = np.vectorize(lambda x: int(x[0]))(array)
ar_gen = np.vectorize(lambda x: 1 if x[4]=='M' else 0)(array)
ar_gen = np.vectorize(lambda x: 1 if x[4]=='M' else 0)(array)
print(ar_id)
print(ar_gen)
def np_vec():
ar_id = np.vectorize(lambda x: int(x[0]))(array)