def main(output_path):
# read building value
if sys.platform == "win32":
working_path = u"R:\scenario_Sydney"
else:
working_path = os.path.join(os.path.expanduser("~"), "Projects/scenario_Sydney")
# output_path = os.path.join(working_path,
# 'scen_risk_parallel')
input_path = os.path.join(working_path, "input")
# file_bldg_value = os.path.join(output_path, site_tag + '_building_value.txt')
# not sure with the first element
# dd = pd.read_csv(file_bldg_value, header=None)
data = read_sitedb_data(os.path.join(input_path, "sitedb_sydney_soil.csv"))
# better with data
# read gmotion
# (_, _, _, mmi) = read_gm(output_path, site_tag='sydney_soil')
# data['MMI'] = pd.Series(mmi[:, 0], index=data.index)
# list building loss txt files
files_loss_txt = glob.glob(os.path.join(output_path, site_tag + "_building_loss.txt*"))
bldg_loss = dict()
for file_ in files_loss_txt:
with open(file_, "r") as fid:
tmp = fid.read().strip().split("\n")
bid = np.array(tmp[1].split(), dtype=np.int64)
loss = np.array(tmp[2].split(), dtype=float)
assert len(bid) == len(loss)
for bid_, loss_ in zip(bid, loss):
if bid_ in bldg_loss:
raise ("Something is wrong")
else:
bldg_loss[bid_] = loss_
# merge to data by bid
df_bldg_loss = pd.DataFrame.from_dict(bldg_loss, orient="index")
df_bldg_loss.reset_index(level=0, inplace=True)
df_bldg_loss = df_bldg_loss.rename(columns={"index": "BID", 0: "LOSS"})
ndata = pd.merge(data, df_bldg_loss, on="BID", how="outer")
# file_bldg_loss = os.path.join(output_path, site_tag + '_building_loss.txt')
# with open(file_bldg_loss, 'r') as fid:
# tmp = fid.read().strip().split('\n')
# bldg_loss = np.array(tmp[2].split(), dtype=float)
ndata["LOSS_RATIO"] = ndata["LOSS"] / data["BLDG_COST"]
def main(args):
output_path, bldg_class = args[0], args[1]
for item in os.listdir(output_path):
if 'event' in item and os.path.isdir(os.path.join(output_path, item)):
site_tag = item.replace('_event_set', '')
# read building value
if sys.platform == 'win32':
working_path = u'R:\scenario_Sydney'
else:
working_path = os.path.join(os.path.expanduser("~"),
'Projects/scenario_Sydney')
file_event = os.path.join(output_path, site_tag + '_event_set',
'event_activity', 'event_activity.npy')
activity = np.load(file_event)[0, 0, 0, :] # (1, 1, 1, nevents)
nevents = len(activity)
input_path = os.path.join(working_path, 'input')
data = read_sitedb_data(os.path.join(input_path,
'sitedb_sydney_soil.csv'))
df_bldg_cost = data.groupby('STRUCTURE_CLASSIFICATION')['BLDG_COST'].sum()
# denominator
if bldg_class.lower() == 'all':
bldg_cost = df_bldg_cost.sum()
else:
try:
bldg_cost = df_bldg_cost[bldg_class]
except KeyError:
print "invlid bldg class: {}".format(bldg_class)
# better with data
# read gmotion
#(_, _, _, mmi) = read_gm(output_path, site_tag='sydney_soil')
#data['MMI'] = pd.Series(mmi[:, 0], index=data.index)
# list building loss txt files
files_loss_txt = glob.glob(os.path.join(output_path,
site_tag + '_building_loss.txt*'))
loss_array = np.zeros((nevents, len(files_loss_txt)))
for i, file_ in enumerate(files_loss_txt):
with open(file_) as fid:
tmp = fid.read().strip().split('\n')
assert len(tmp[2:]) == nevents
loss_ = np.array(tmp[2:], dtype=float)/bldg_cost
loss_array[:, i] = loss_[:]
loss_array = np.sum(loss_array, axis=1)
result = np.vstack((activity, loss_array)).T
np.save(os.path.join(output_path,
'activity_loss_{}.npy'.format(bldg_class)), result)
return result
