def prepare_edge(edge_shapefile, building_shapefile):
"""Create edge graph with grouped building demands.
"""
# load buildings and sum by type and nearest edge ID
# 1. read shapefile to DataFrame (with special geometry column)
# 2. group DataFrame by columns 'nearest' (ID of nearest edge) and 'type'
# (residential, commercial, industrial, other)
# 3. sum by group and unstack, i.e. convert secondary index 'type' to columns
buildings = pdshp.read_shp(building_shapefile)
building_type_mapping = {
'church': 'other',
'farm': 'other',
'hospital': 'residential',
'hotel': 'commercial',
'house': 'residential',
'office': 'commercial',
'retail': 'commercial',
'school': 'commercial',
'yes': 'other'
}
buildings.replace(to_replace={'type': building_type_mapping}, inplace=True)
buildings_grouped = buildings.groupby(['nearest', 'type'])
total_area = buildings_grouped.sum()['AREA'].unstack()
# load edges (streets) and join with summed areas
# 1. read shapefile to DataFrame (with geometry column)
# 2. join DataFrame total_area on index (=ID)
# 3. fill missing values with 0
edge = pdshp.read_shp(edge_shapefile)
edge = edge.set_index('Edge')
edge = edge.join(total_area)
edge = edge.fillna(0)
return edge
def prepare_edge(edge_shapefile, building_shapefile):
"""Create edge graph with grouped building demands.
"""
# load buildings and sum by type and nearest edge ID
# 1. read shapefile to DataFrame (with special geometry column)
# 2. group DataFrame by columns 'nearest' (ID of nearest edge) and 'type'
# (residential, commercial, industrial, other)
# 3. sum by group and unstack, i.e. convert secondary index 'type' to
# columns
buildings = geopandas.read_file(building_shapefile + '.shp')
buildings = buildings.convert_objects(convert_numeric=True)
building_type_mapping = {
'basin': 'other', 'chapel': 'other', 'church': 'other',
'farm_auxiliary': 'other', 'greenhouse': 'other',
'school': 'public',
'office': 'commercial', 'restaurant': 'commercial',
'yes': 'residential', 'house': 'residential'}
buildings.replace(to_replace={'type': building_type_mapping}, inplace=True)
buildings = buildings.to_crs(epsg=32632)
buildings['AREA'] = buildings.area
buildings_grouped = buildings.groupby(['nearest', 'type'])
total_area = buildings_grouped.sum()['AREA'].unstack()
# load edges (streets) and join with summed areas
# 1. read shapefile to DataFrame (with geometry column)
# 2. join DataFrame total_area on index (=ID)
# 3. fill missing values with 0
edge = pdshp.read_shp(edge_shapefile)
edge = edge.set_index('Edge')
edge = edge.join(total_area)
edge = edge.fillna(0)
return edge
def run_scenario(scenario, result_dir):
# scenario name
sce = scenario.__name__
sce_nice_name = sce.replace('_', ' ').title()
# prepare input data
data = rivus.read_excel(data_spreadsheet)
vertex = pdshp.read_shp(vertex_shapefile)
edge = prepare_edge(edge_shapefile, building_shapefile)
# apply scenario function to input data
data, vertex, edge = scenario(data, vertex, edge)
log_filename = os.path.join(result_dir, sce+'.log')
# create & solve model
prob = rivus.create_model(
data, vertex, edge,
peak_multiplier=lambda x:scale_peak_demand(x, peak_demand_prefactor))
# scale peak demand according to pickled urbs findings
#reduced_peak = scale_peak_demand(model, peak_demand_prefactor)
#model.peak = reduced_peak
if PYOMO3:
prob = prob.create()
optim = SolverFactory('glpk')
optim = setup_solver(optim, logfile=log_filename)
result = optim.solve(prob, tee=True)
if PYOMO3:
prob.load(result)
# report
rivus.save(prob, os.path.join(result_dir, sce+'.pgz'))
rivus.report(prob, os.path.join(result_dir, sce+'.xlsx'))
# plot without buildings
rivus.result_figures(prob, os.path.join(result_dir, sce))
# plot with buildings and to_edge lines
more_shapefiles = [{'name': 'to_edge',
'color': rivus.to_rgb(192, 192, 192),
'shapefile': to_edge_shapefile,
'zorder': 1,
'linewidth': 0.1}]
rivus.result_figures(prob, os.path.join(result_dir, sce+'_bld'),
buildings=(building_shapefile, False),
shapefiles=more_shapefiles)
return prob
def run_scenario(scenario):
# scenario name
sce = scenario.__name__
sce_nice_name = sce.replace('_', ' ').title()
# prepare input data
data = rivus.read_excel(data_spreadsheet)
vertex = pdshp.read_shp(vertex_shapefile)
edge = prepare_edge(edge_shapefile, building_shapefile)
# apply scenario function to input data
data, vertex, edge = scenario(data, vertex, edge)
# create & solve model
prob = rivus.create_model(data, vertex, edge)
if PYOMO3:
prob = prob.create() # no longer needed in Pyomo 4+
optim = SolverFactory('gurobi')
optim = setup_solver(optim)
result = optim.solve(prob, tee=True)
if PYOMO3:
prob.load(result) # no longer needed in Pyomo 4+
# create result directory if not existent
result_dir = os.path.join('result', os.path.basename(base_directory))
if not os.path.exists(result_dir):
os.makedirs(result_dir)
# report
rivus.report(prob, os.path.join(result_dir, 'report.xlsx'))
# plots
for com, plot_type in [('Elec', 'caps'), ('Heat', 'caps'), ('Gas', 'caps'),
('Elec', 'peak'), ('Heat', 'peak')]:
# two plot variants
for plot_annotations in [False, True]:
# create plot
fig = rivus.plot(prob,
com,
mapscale=False,
tick_labels=False,
plot_demand=(plot_type == 'peak'),
annotations=plot_annotations)
plt.title('')
# save to file
for ext, transp in [('png', True), ('png', False), ('pdf', True)]:
transp_str = ('-transp' if transp and ext != 'pdf' else '')
annote_str = ('-annote' if plot_annotations else '')
# determine figure filename from scenario name, plot type,
# commodity, transparency, annotations and extension
fig_filename = '{}-{}-{}{}{}.{}'.format(
sce, plot_type, com, transp_str, annote_str, ext)
fig_filename = os.path.join(result_dir, fig_filename)
fig.savefig(fig_filename,
dpi=300,
bbox_inches='tight',
transparent=transp)
return prob
# reference with list of options
# execute 'glpsol --help'
optim.set_options("tmlim=600")
optim.set_options("mipgap=2e-2")
else:
print("Warning from setup_solver: no options set for solver "
"'{}'!".format(optim.name))
return optim
# load buildings and sum by type and nearest edge ID
# 1. read shapefile to DataFrame (with special geometry column)
# 2. group DataFrame by columns 'nearest' (ID of nearest edge) and 'type'
# (residential, commercial, industrial, other)
# 3. sum by group and unstack, i.e. convert secondary index 'type' to columns
buildings = pdshp.read_shp(building_shapefile)
building_type_mapping = {
'church': 'other',
'farm': 'other',
'hospital': 'residential',
'hotel': 'commercial',
'house': 'residential',
'office': 'commercial',
'retail': 'commercial',
'school': 'commercial',
'yes': 'other',
}
buildings.replace(to_replace={'type': building_type_mapping}, inplace=True)
buildings_grouped = buildings.groupby(['nearest', 'type'])
total_area = buildings_grouped.sum()['AREA'].unstack()