Homec4science
Diffusion solar_potential (aw_solar) Browse

solarPOT/DOM/merge_roofs_and_buildings.py
c54246b01dfdaw_solar

Last Change
Raw File

merge_roofs_and_buildings.py
View Options

import numpy as np
import pandas as pd
import os
import sys
import scipy.spatial as spatial
import util
##=============== FUNCTION DEFINITIONS
def find_neighbors(input_table, radius, x = 'GKODX', y = 'GKODY', ID_column = 'GWR_EGID'):
# create unique non-NaN rows for correct neighbor count
reduced_table = input_table.loc[ : , [ ID_column, x, y ] ].dropna().drop_duplicates().reset_index()
# declare a tree structure to store points and count neighbors (based on 2D coordinates)
tree = spatial.KDTree(reduced_table.loc[:,[x, y]])
# compute all pairs in tree
tt = util.Timer()
query_pairs = tree.query_pairs(r = radius)
print('Finished computing neighboring pairs')
tt.stop()
# convert output to pandas dataframe
query_array = list(query_pairs)
df = pd.DataFrame(np.asarray(query_array), columns = ['index_x', 'index_y'])
# COUNT TOTAL NUMBER OF NEIGHBORS
# group by index, to count number of neighbors per pair (for left-side merging)
left_groups = df.groupby('index_x').count()
left_groups = left_groups.rename({'index_y' : 'n_neighbors'}, axis = 1)
# group by index, to count number of neighbors per pair (for right-side merging)
right_groups = df.groupby('index_y').count()
right_groups = right_groups.rename({'index_x' : 'n_neighbors'}, axis = 1)
# merge counts together and evaluate total number of neighbors for each building
neighbors = left_groups.merge(right_groups, left_index = True, right_index = True, how='outer').fillna(0)
neighbors[('n_neighbors_%d' %radius)] = neighbors.n_neighbors_x + neighbors.n_neighbors_y
all_neighbors = reduced_table.merge(neighbors, left_index = True, right_index = True, how='outer').fillna(0)
# return neighbor count data, merged onto the input table as column n_neighbors_RADIUS
return input_table.merge(all_neighbors, on = [ ID_column, x, y ], how = 'left').drop(['index', 'n_neighbors_x', 'n_neighbors_y'], axis=1)
##================================================
##=========== USER INPUT =========================
# Filenames
SONNENDACH_FILE = sys.argv[1]
GWR_FILE = sys.argv[2]
GWR_COL_FILE = sys.argv[3]
TARGET_FILENAME = sys.argv[4]
# Desired columns
SONNENDACH_COLUMNS = ['DF_UID','FLAECHE','NEIGUNG','AUSRICHTUN','XCOORD','YCOORD', 'SHAPE_Leng', 'SHAPE_Area','GWR_EGID']
GWR_COLUMNS = ['EGID', 'GBAUP', 'GKAT', 'GAREA', 'GASTW', 'GKODX', 'GKODY']
# Contants
NEIGHBOR_RADIUS = 100
##=========== ROOFTOP DATA =======================
# read rooftop data from file
rooftops = pd.read_csv(SONNENDACH_FILE, index_col = 0)
rooftops = rooftops.loc[:, SONNENDACH_COLUMNS ]
# set flat roof's orientation ("AUSRICHTUN") to 0
rooftops.loc[rooftops.NEIGUNG == 0, 'AUSRICHTUN'] = 0
print('Finished loading rooftops')
##=============== GWR DATA =======================
# read GWR data and select relevant columns
GWR_column_names = pd.read_csv( GWR_COL_FILE ).columns
buildings = pd.read_csv(GWR_FILE, sep = '\t', encoding = 'cp1250', header = None, names = GWR_column_names)
GWR = buildings.loc[:,GWR_COLUMNS]
print('Finished loading GWR')
##=== MERGE DATA AND ADD ADDITIONAL VARIABLES ====
all_rooftops = rooftops.merge(GWR, left_on = 'GWR_EGID', right_on = 'EGID', how = 'left').drop('EGID', axis = 1)
all_rooftops['SHAPE_Ratio'] = all_rooftops.SHAPE_Leng/all_rooftops.FLAECHE
print('Merged rooftops and GWR and computed SHAPE_Ratio')
rooftop_info = find_neighbors( all_rooftops, radius = NEIGHBOR_RADIUS )
print('Computed number of neighbors')
##======= Save as csv ============================
rooftop_info.to_csv(TARGET_FILENAME)
print('Saved output table in %s' %TARGET_FILENAME)
c4science · Help