class StoreGenerator(object):
def __init__(self, zipcode_pop):
self.zipcode_pop = zipcode_pop
pdf = ZipcodePDF(zipcode_pop)
generator = ZipcodeGenerator(zipcode_pop)
self.mcmc = MonteCarloGenerator(generator, pdf)
def generate_n(self, n):
zipcodes = self.generate_zipcodes(n)
stores = self.create_stores(zipcodes)
return stores
def generate_zipcodes(self, n):
zipcodes = []
for i in xrange(n):
zipcode = self.mcmc.generate()
zipcodes.append(zipcode)
return zipcodes
def create_stores(self, zipcodes):
ids = np.arange(len(zipcodes))
populations = self.zipcode_pop.loc[zipcodes]
total_pop = np.float64(populations.sum(axis=0))
pop_density = populations.divide(total_pop)
return pd.DataFrame(data={"id" : ids, "zipcode" : zipcodes, "pop_density" : pop_density}, index=zipcodes)
class CustomerGenerator(object):
def __init__(self, stores):
self.stores = stores
gen = CustomerInstanceGenerator(stores)
pdf = CustomerPDF(stores)
self.mcmc = MonteCarloGenerator(gen, pdf)
def generate_n(self, n):
prototypes = self.generate_customers(n)
finalized = self.finalize_customers(prototypes)
return finalized
def generate_customers(self, n):
customers = []
for i in xrange(n):
customer = self.mcmc.generate()
customers.append(customer)
return customers
def finalize_customers(self, customers):
zipcodes = []
pets = []
names = []
purchase_values = []
value_total = 0.0
ids = np.arange(len(customers))
for c in customers:
zipcodes.append(c["zipcode"])
pets.append(c["pet_type"])
names.append(c["name"])
purchase_values.append(c["purchase_value"])
value_total += c["purchase_value"]
purchase_pdf = [value / value_total for value in purchase_values]
return pd.DataFrame(data={"id" : ids, "name" : names, "zipcode" : zipcodes, "pet_type" : pets, "purchase_pdf" : purchase_pdf}, index=ids)
class StoreSampler(object): def __init__(self, stores): pdf = StoreSamplerPDF(stores) gen = StoreSamplerGenerator(stores) self.mcmc = MonteCarloGenerator(gen, pdf) def sample(self): return self.mcmc.generate()
class CustomerGenerator(object):
def __init__(self, stores):
self.stores = stores
gen = CustomerInstanceGenerator(stores)
pdf = CustomerPDF(stores)
self.mcmc = MonteCarloGenerator(gen, pdf)
def generate_n(self, n):
prototypes = self.generate_customers(n)
finalized = self.finalize_customers(prototypes)
return finalized
def generate_customers(self, n):
customers = []
for i in xrange(n):
customer = self.mcmc.generate()
customers.append(customer)
return customers
def finalize_customers(self, customers):
zipcodes = []
pets = []
names = []
purchase_values = []
value_total = 0.0
ids = np.arange(len(customers))
for c in customers:
zipcodes.append(c["zipcode"])
pets.append(c["pet_type"])
names.append(c["name"])
purchase_values.append(c["purchase_value"])
value_total += c["purchase_value"]
purchase_pdf = [value / value_total for value in purchase_values]
return pd.DataFrame(data={
"id": ids,
"name": names,
"zipcode": zipcodes,
"pet_type": pets,
"purchase_pdf": purchase_pdf
},
index=ids)
def __init__(self, stores):
self.stores = stores
gen = CustomerInstanceGenerator(stores)
pdf = CustomerPDF(stores)
self.mcmc = MonteCarloGenerator(gen, pdf)
def __init__(self, stores): pdf = StoreSamplerPDF(stores) gen = StoreSamplerGenerator(stores) self.mcmc = MonteCarloGenerator(gen, pdf)
def __init__(self, zipcode_pop): self.zipcode_pop = zipcode_pop pdf = ZipcodePDF(zipcode_pop) generator = ZipcodeGenerator(zipcode_pop) self.mcmc = MonteCarloGenerator(generator, pdf)
def __init__(self, stores): self.stores = stores gen = CustomerInstanceGenerator(stores) pdf = CustomerPDF(stores) self.mcmc = MonteCarloGenerator(gen, pdf)
