def choose_candidate(candidates, mgras, product_type_labels, max_units):
# Filter
filtered = apply_filters(candidates, product_type_labels)
if len(filtered) < 1:
print('out of suitable mgras for product type {}'.format(
product_type_labels.product_type))
print('evaluate filtering methods...\nexiting')
return None
# !
# save_to_file(filtered, 'data/output', 'filtered_candidates.csv')
# return
# Sample
weights = combine_weights(filtered.profit_margin, filtered.vacancy_cap)
selected_candidate = filtered.sample(n=1, weights=weights)
buildable_count = buildable_units(selected_candidate, product_type_labels,
max_units)
# develop buildable_count units by updating the MGRA in the original
# dataframe
removed_units_reference = update_mgra(mgras,
selected_candidate,
buildable_count,
product_type_labels,
scheduled_development=False)
return mgras, buildable_count, removed_units_reference
def test_update_mgra(self):
# print(self.test_mgras.iloc[1223, :])
random_mgra = self.test_mgras.sample(random_state=19)
self.assertIsNone(
update_mgra(self.test_mgras,
random_mgra,
10,
self.product_type_labels,
scheduled_development=False))
def test_update_mgra(self):
random_mgra = self.test_mgras.sample(random_state=19)
self.assertIsNone(
update_mgra(self.test_mgras,
random_mgra,
10,
self.product_type_labels,
scheduled_development=False))
for label in self.vacant_labels:
self.assertFalse(any(self.test_mgras[label] < 0))
def add_to_mgra(mgras, site):
'''
mgras: the dataframe to update.
site: expects a dataframe with one entry, including columns for MGRAID
and the number of units to add
'''
mgra_id = site.MGRA.item()
single_family_units = site.sfu.item()
multi_family_units = site.mfu.item()
employment_units = site.civEmp.item()
if single_family_units > 0:
# add single family units
labels = ProductTypeLabels(SINGLE_FAMILY)
update_mgra(mgras, mgra_id, single_family_units, labels)
reduce_demand(labels.product_type, single_family_units)
if multi_family_units > 0:
# add multifamily units
labels = ProductTypeLabels(MULTI_FAMILY)
update_mgra(mgras, mgra_id, multi_family_units, labels)
reduce_demand(labels.product_type, multi_family_units)
if employment_units > 0:
# add employment; determine which type to add
mgra = mgras.loc[mgras[mgra_labels.MGRA] == mgra_id]
labels = find_current_employment_type(mgras, mgra)
update_mgra(mgras, mgra_id, employment_units, labels)
reduce_demand(labels.product_type, employment_units)
return
def choose_candidate(candidates, mgras, product_type_labels, max_units):
selected_candidate = candidates.select_candidate_for_product_type(
product_type_labels.product_type)
if selected_candidate is None:
logging.error(
'out of suitable mgra candidates for product type {}'.format(
product_type_labels.product_type))
logging.error('evaluate input data and/or filtering methods...')
logging.error('setting remaining demand for {} to zero'.format(
product_type_labels.product_type))
return None, None
logging.debug('selected candidate: {}'.format(selected_candidate))
buildable_count = buildable_units(
selected_candidate, product_type_labels, max_units)
if buildable_count < 1:
return 0, None
# develop buildable_count units on the selected MGRA by updating it in the
# original dataframe
removed_units_reference = update_mgra(
mgras, selected_candidate,
buildable_count, product_type_labels,
scheduled_development=False, candidates=candidates.candidates)
return buildable_count, removed_units_reference