!pip install --q Faker

import string
from faker import Faker 
import json                 
import numpy as np
import pandas as pd
fake = Faker() 

fake = Faker('pt_BR')

fake_name=[]
for i in range(10000):
  #print("Name:", i)
  name = fake.name()
  #print(name)
  fake_name.append(name)

len(fake_name)

lista_unic = set(fake_name)

len(lista_unic)

# Total lines created
n = 1000

infos = ['name','job', 'company','ssn', 'address', 'sex', 'birthdate']

def create_profile(x): 
    print("Creating profile with ", n, " lines")
    # dictionary 
    profile_data ={} 
    for i in range(0, x): 
        profile_data[i]={} 
        profile_data[i] = fake.profile(infos)

    print("done")
    return profile_data
    


def create_sales(x): 
    print("Creating sales with ", n, " lines")
  
    # dictionary 
    sales_data ={} 
    for i in range(0, x): 
        sales_data[i]={} 
        sales_data[i]['price'] = np.random.randint(10,500)
        sales_data[i]['qtd'] = np.random.randint(1,10)
        sales_data[i]['product'] =  chr(np.random.randint(ord('A'), ord('Z')))
        sales_data[i]['bill'] = sales_data[i]['price'] * sales_data[i]['qtd']
        sales_data[i]['way_of_payment'] = np.random.choice(['boleto', 'cartão_cred', 'a vista'])
    print("done")
    return sales_data


# Create fake profile
profiles = create_profile(n)
df_prof = pd.DataFrame.from_dict(profiles)
df_profile = df_prof.T

# Create fake sales
sales = create_sales(n)
temp_sales = pd.DataFrame.from_dict(sales)
df_sale = temp_sales.T

df_fake = pd.concat([df_profile, df_sale], axis=1)

# Extracting and creating column UF
lista_uf=[]
for i in range(len(df_fake['address'])):
  lista_uf.append(df_fake['address'][i].split(' / ')[1])

df_fake['uf'] = lista_uf
df_fake.drop_duplicates(inplace=True)
df_fake