Page MenuHomec4science
Files F119639697

get_ncorners.py
No OneTemporary
Actions

File Metadata

Created
Sat, Jun 28, 04:58

get_ncorners.py
View Options

import numpy as np
import pandas as pd
import geopandas as gpd
from shapely.geometry import Polygon
import fiona
import geo_ops
import os
import glob
import sys
import time
################# USER INPUTS ################
# file information
ROOFS_FILE = sys.argv[1] # NOTE: must be a shapefile
OUTFILE = sys.argv[2]
# Current batch information
BATCH_SIZE = int(sys.argv[3])
################ CONSTANTS ###################
# field names in shapefile
AREA_FIELD = 'FLAECHE'
TILT_FIELD = 'NEIGUNG'
ORIENTATION_FIELD = 'AUSRICHTUN'
# N_ITER = int(np.ceil( float(FILESIZE) / float(BATCH_SIZE) ))
TOLERANCE = 0.1
#============================================#
#============================================#
############### Main function ###############
def main():
print_inputs()
timer = time.time()
corner_count = []
# open file with fiona and get file length / number of iterations
file_handle = fiona.open(ROOFS_FILE)
N_ITER = int(np.ceil( len(file_handle) / float(BATCH_SIZE) ))
print('Length of file: %d' %len(file_handle))
# load data one at a time
for BATCH_ID in range(N_ITER):
print('Starting iteration %d after %.3f seconds' %( BATCH_ID+1, time.time()-timer ))
# load data and perform pre-processing steps (clean & rotate)
roofs = load_from_file(file_handle, batch_id = BATCH_ID, batch_size = BATCH_SIZE, tolerance = TOLERANCE)
# count the numbers of corners of each roof shape
corner_count.append( get_n_corners(roofs) )
corners_all = pd.concat(corner_count)
corners_all.to_csv(OUTFILE)
print('Saved file with n_corner information to %s' %OUTFILE)
file_handle.close()
################ SUBROUTINES ################
def load_from_file(file, batch_id, batch_size, tolerance):
# load a given batch of lines from file and pre-process the data
# returns geopandas dataframe with polygons and database information
# load selected rows from file
start = batch_id * batch_size
end = (batch_id + 1) * batch_size
roofs = gpd.GeoDataFrame.from_features(file[start : end])
if len(roofs) == 0:
raise ValueError('Rooftop dataframe is empty - no data loaded from file.')
# PRE-PROCESSING: simplify roof shapes
roofs['geometry'] = roofs.simplify(tolerance = tolerance)
return roofs
def get_n_corners(roofs):
# compute the number of corners for each geometry in roofs
# output: dataframe with fields "DF_UID" and "n_corners", can then be merged with rooftops
# step 1: explode roof shapes
unique_gdf = roofs.explode().reset_index().drop(['level_0', 'level_1'], axis = 1)
# step 2: attach number of corners to the unique dataframe
# (take len - 1 as connection point of rectangle is always counted twice)
try:
unique_gdf['n_corners'] = unique_gdf.geometry.exterior.apply(lambda x: len(x.coords)-1)
except:
print('Adding n_corners failed: Likely due to failure because of multipart polygons')
# step 3: drop unnecessary data
corner_df = unique_gdf.loc[:, ['DF_UID', 'n_corners']]
# step 4: merge by DF_UID and get the count (sum over all roof pieces)
return corner_df.groupby('DF_UID').sum()
############ AUXILARY FUNCTIONS ##############
def print_inputs():
print('\nINPUT DATA:')
print("Input file: %s" %ROOFS_FILE)
print("Output file: %s" %OUTFILE)
print("Batch size: %d" %BATCH_SIZE)
print("")
##############################################
if __name__ == '__main__':
main()

Event Timeline

c4science · Help