Adding update_parking script back in.

run_soundcast.py

@@ -54,7 +54,10 @@ def parcel_buffering():
         else:
             print 'Starting to update UrbanSim parcel data with 4k parking data file'
             returncode = subprocess.call([sys.executable,
-                                      'scripts/utils/ParcelBuffering/update_parking.py', base_inputs])
+                                      'scripts/utils/update_parking.py', base_inputs])
+            if returncode != 0:
+                print 'Update Parking failed'
+                sys.exit(1)
             print 'Finished updating parking data on parcel file'
 
     create_buffer_xml()

scripts/utils/update_parking.py

@@ -0,0 +1,95 @@
+# This script modifies parcel-level parking data. 
+# Parcels obtain aggregate parking costs from 4K ensembles. 
+# Parcels with parking cost but zero parking spaces are given the average number of parking spaces. 
+# Currently this average is for the entire region, but should be updated to properly model number of spaces
+# by building type, location, or local code. This is just a placeholder to remove coding errors in DaySim
+# when zero-parking space parcels cause divide-by-zero errors. 
+#
+# Run this script from the same location as the parcels.txt file
+# Parcels.txt can be converted from .dbf format with ArcGIS.
+
+import pandas as pd
+import h5py
+import numpy as np
+from input_configuration import *
+
+
+daily_parking_cost = "inputs\\parcel_buffer\\daily_parking_costs.csv"
+hourly_parking_cost = "inputs\parcel_buffer\\hourly_parking_costs.csv"
+input_ensemble = "inputs\\parking_gz.csv"
+input_parcels = "inputs\\parcel_buffer\\parcels_urbansim.txt"
+
+# Combine data columns
+df_parcels = pd.read_csv(input_parcels, delim_whitespace=True)
+df_ensemble = pd.read_csv(input_ensemble, low_memory = False)
+df_daily_parking_cost = pd.DataFrame(pd.read_csv(daily_parking_cost, low_memory = False))
+df_hourly_parking_cost = pd.DataFrame(pd.read_csv(hourly_parking_cost, low_memory = False))
+join_ensemble_to_parcel = pd.merge(left = df_parcels, right=df_ensemble,left_on="TAZ_P",right_on="TAZ")
+# Join daily costs with parcel data 
+merged_df = pd.merge(left=join_ensemble_to_parcel,
+                                right=df_daily_parking_cost, 
+                                left_on="ENS",
+                                right_on="ENS")
+# Join hourly costs with parcel data
+merged_df = pd.merge(left = merged_df,
+                                right = df_hourly_parking_cost,
+                                left_on="ENS",
+                                right_on="ENS")
+
+
+# Clean up the results and store in same format as original parcel.txt file
+df = pd.DataFrame(merged_df)
+#drop_columns = ['PARKDY_P', 'PARKHR_P', 'PPRICDYP', 'PPRICHRP']
+drop_columns = ['PPRICDYP', 'PPRICHRP']
+df = df.drop(drop_columns, 1)
+
+for column_title in drop_columns:
+    df = df.rename(columns = {'DAY_COST':'PPRICDYP',
+                              'HR_COST':'PPRICHRP'})
+
+
+# For parcels in regions with non-zero parking costs, ensure each parcel has some minumum parking spaces available
+# For now, replacing all zero-parking locations with region-wide average number of parking spots. 
+# We can update this with an ensemble-wide average from GIS analysis later.
+#
+daily_parking_zeros = df[df['PPRICDYP'] > 0]['PPRICDYP']
+hourly_parking_zeros = df[df['PPRICHRP'] > 0]['PPRICHRP']
+avg_daily_parking = int(daily_parking_zeros.mean())
+avg_hourly_parking = int(hourly_parking_zeros.mean())
+
+replace_daily_parking_costs = daily_parking_zeros[daily_parking_zeros == 0]
+replace_hourly_parking_costs = hourly_parking_zeros[hourly_parking_zeros == 0]
+f_dy = lambda x : avg_daily_parking
+f_hr = lambda x : avg_hourly_parking
+replace_daily_parking_costs = replace_daily_parking_costs.apply(f_dy)
+replace_daily_parking_costs = pd.DataFrame(replace_daily_parking_costs)
+replace_hourly_parking_costs = replace_hourly_parking_costs.apply(f_hr)
+replace_hourly_parking_costs = pd.DataFrame(replace_hourly_parking_costs)
+
+# merge results back in to main data
+if len(replace_daily_parking_costs) > 0:
+    df = df.join(replace_daily_parking_costs, lsuffix = '_left', rsuffix = '_right')
+    df = df.rename(columns = {'PARKDY_P_right':'PARKDY_P'})
+    if 'PARKDY_P_left' in df:
+        df = df.drop('PARKDY_P_left', 1)
+    df = df.join(replace_hourly_parking_costs, lsuffix = '_left', rsuffix = '_right')
+    df = df.rename(columns = {'PARKHR_P_right':'PARKHR_P'})
+    if 'PARKHR_P_left' in df:
+        df = df.drop('PARKHR_P_left', 1)
+
+# check if any lines have zero costs and more than zero paid parking spaces
+#check_day = df[df['PPRICDYP'] > 0]['PARKDY_P']
+#check_hour = df[df['PPRICHRP'] > 0]['PARKHR_P']
+#if check_day.sum()/check_day.count() != check_day.mean():
+#    print "Warning. Some parcels may have zero daily parking spaces but non-zero price."
+#if check_hour.sum()/check_hour.count() != check_hour.mean():
+#    print "Warning. Some parcels may have zero hourly parking spaces but non-zero price."
+
+# Delete TAZ and ENS columns to restore file to original form. 
+df = df.drop(['ENS', 'TAZ'], 1)
+
+# Save results to text file
+df.to_csv(input_parcels, sep = ' ', index = False)
+
+# End the script
+print "Parcel file updated with aggregate parking costs and lot numbers. " + str(len(replace_daily_parking_costs)) + " parcels were updated."