import pandas as pd
from plots.packing_algo import ConvexPolygon, pack_polygons, truncate, RectangularContainer, ConvexContainer
from plots.polygon_creator import voronoi_polygons_wrapper, rectangle_cutter
def build_dataset_with_rectangle_cutter(rect_width: float, rect_height, repetition, cut_list=[], angle_steps=90,
cut_min=1,
cut_max=1, cut_steps=1, intervals=[0, 0.01, 0.05, 1], weights=[0, 0, 0.5, 1],
save_container=True, create_plots=True):
cut_numbers = []
data_dict_list = []
if cut_list:
cut_numbers = cut_list
else:
if cut_min > cut_max:
cut_max = cut_min
for cut_number in range(cut_min, cut_max + 1, cut_steps):
cut_numbers.append(cut_number)
for cut_count in cut_numbers:
cutted_polygons_lists = []
for n in range(0, repetition):
cutted_polygons = rectangle_cutter(rect_width, rect_height, cut_count, intervals=intervals, weights=weights)
cutted_polygons_lists.append(cutted_polygons)
dict_data = collect_rect_containers_data(rect_width, rect_height, cutted_polygons_lists, angle_steps,
save_container, create_plots=True)
data_dict_list.append(dict_data)
return data_dict_list
def build_dataset_with_voronoi(rect_width, rect_height, repetition, cut_list=[], cut_min=5, cut_max=5, cut_steps=1,
angle_steps=90, save_container=True, create_plots=True):
cut_numbers = []
data_dict_list = []
if cut_list:
cut_numbers = cut_list
else:
if cut_min > cut_max:
cut_max = cut_min
for cut_number in range(cut_min, cut_max + 1, cut_steps):
cut_numbers.append(cut_number)
for cut_count in cut_numbers:
cutted_polygons_lists = []
for n in range(0, repetition):
cutted_polygons = voronoi_polygons_wrapper(rect_width, rect_height, cut_count)
cutted_polygons_lists.append(cutted_polygons)
dict_data = collect_rect_containers_data(rect_width, rect_height, cutted_polygons_lists, angle_steps,
save_container, create_plots=True)
data_dict_list.append(dict_data)
return data_dict_list
def collect_rect_containers_data(rect_width: float, rect_height: float, cutted_polygons_lists: [[ConvexPolygon]],
angle_steps=90, save_container=True, create_plots=True) -> {str, list}:
opt_area = rect_width * rect_height
opt_area_list = []
area_list = []
area_div_list = []
angle_0_area_div_list = []
angle_0_not_clipped_area_div_list = []
not_clipped_area_list = []
not_clipped_area_div_list = []
polygon_count_list = []
end_container_list_plots = []
for polygons in cutted_polygons_lists:
cc = ConvexContainer(polygons, steps=angle_steps, build_plots=create_plots)
end_container = cc.smallest_rectangular_container
end_container_angle_0 = cc.angle_0_rectangular_container
c_area = end_container.container_area
c_not_opt_area = end_container.container_not_clipped_area
c_not_r_area = end_container_angle_0.container_area
c_not_r_not_opt_area = end_container_angle_0.container_not_clipped_area
area_list.append(c_area)
area_div_list.append(c_area / opt_area)
not_clipped_area_list.append(c_not_opt_area)
not_clipped_area_div_list.append(c_not_opt_area / opt_area)
angle_0_area_div_list.append(c_not_r_area / opt_area)
angle_0_not_clipped_area_div_list.append(c_not_r_not_opt_area / opt_area)
opt_area_list.append(opt_area)
polygon_count_list.append(len(polygons))
if save_container:
end_container_list_plots.append(cc.plot_container(render=False))
rect_containers_data_dict = {'area': area_list, 'area/opt_area': area_div_list,
'not_clipped_area': not_clipped_area_list,
'not_clipped_area/opt_area': not_clipped_area_div_list,
'angle_0_area/opt_area': angle_0_area_div_list,
'angle_0_not_clipped_area/opt_area': angle_0_not_clipped_area_div_list,
'opt-area': opt_area_list,
'polygon count': polygon_count_list}
if end_container_list_plots:
rect_containers_data_dict['End-Container'] = end_container_list_plots
return rect_containers_data_dict
# # need to update like rectangle_cutter with deviation
# def find_repition_factor_rectangle_voronoi(cut, rep_high, rep_low=1, rep_steps=1, display_flag=True,
# intervals=[0, 0.01, 0.05, 1], weights=[0, 0, 0.5, 1]):
# average_list = []
# if rep_low < 1 or rep_high < 1:
# raise ValueError("the value of rep_high and rep_low need to bigger then 0")
# for rep in range(rep_low, rep_high + rep_steps, rep_steps):
# data = build_dataset_with_voronoi(1000, 1000, rep, cut_min=cut, cut_max=cut, cut_steps=1)
# df = pd.DataFrame(data[0])
# if display_flag:
# display(df.sort_values(by="area", ascending=False))
# list_optimal_areas = df["area/opt_area"].tolist()
# average = sum(list_optimal_areas) / len(list_optimal_areas)
# average_list.append((rep, average))
# return average_list
# # deviation=0.05, accept=10
# def find_repition_factor_rectangle_cutter(cut, rep_high, deviation=0.05, accept_number=10, rep_low=1, rep_steps=1,
# display_flag=True, intervals=[0, 0.01, 0.05, 1], weights=[0, 0, 0.5, 1]):
# average_list = []
# if rep_low < 1 or rep_high < 1:
# raise ValueError("the value of rep_high and rep_low need to bigger then 0")
# average_canidate = 0
# accept_counter = accept_number
# average_top_bound = average_canidate + average_canidate * deviation
# average_bot_bound = average_canidate - average_canidate * deviation
# for rep in range(rep_low, rep_high + rep_steps, rep_steps):
# data = build_dataset_with_rectangle_cutter(1000, 1000, rep, cut_min=cut, cut_max=cut, cut_steps=1,
# intervals=[0, 0.01, 0.05, 1], weights=[0, 0, 0.5, 1])
# df = pd.DataFrame(data[0])
# if display_flag:
# display(df.sort_values(by="area", ascending=False))
# list_optimal_areas = df["area/opt_area"].tolist()
# average = sum(list_optimal_areas) / len(list_optimal_areas)
# if average_bot_bound <= average <= average_top_bound:
# accept_counter -= 1
# else:
# average_canidate = average
# accept_counter = accept_number
# average_top_bound = average_canidate + average_canidate * deviation
# average_bot_bound = average_canidate - average_canidate * deviation
# if accept_counter <= 0:
# print(rep)
# return average_list
# average_list.append((rep, average))
# return average_list
def dict_list_to_ordered_panda_list(dictionary_list: [dict], ordered_by_area=True) -> [pd.DataFrame]:
data_frame_list = []
for dic in dictionary_list:
df = pd.DataFrame(dic)
if ordered_by_area:
data_frame_list.append(df.sort_values(by="area", ascending=False))
else:
data_frame_list.append(df.sort_values(by="not_clipped_area", ascending=False))
return data_frame_list
# def display_panda_df_list(df_list: [pd.DataFrame]) -> None:
# for df in df_list:
# display(df)
def build_aprox_values(panda_data):
mean_opt = 0
mean_not_clipped = 0
mean_angle_0 = 0
mean_angle_0_not_clipped = 0
counter = 0
mean_best = 0
mean_worst = 0
for data in panda_data:
mean_opt += data["area/opt_area"].mean(axis=0)
mean_best += data["area/opt_area"].iloc[-1]
mean_worst += data["area/opt_area"].iloc[0]
mean_not_clipped += data["not_clipped_area/opt_area"].mean(axis=0)
mean_angle_0 += data["angle_0_area/opt_area"].mean(axis=0)
mean_angle_0_not_clipped += data["angle_0_not_clipped_area/opt_area"].mean(axis=0)
counter += 1
mean_opt = mean_opt / counter
mean_best = mean_best / counter
mean_worst = mean_worst / counter
mean_not_clipped = mean_not_clipped / counter
mean_angle_0 = mean_angle_0 / counter
mean_angle_0_not_clipped = mean_angle_0_not_clipped / counter
aprox_dict = dict(aprox=mean_opt, aprox_best=mean_best, aprox_worst=mean_worst, aprox_not_clipped=mean_not_clipped,
aprox_angle_0=mean_angle_0, aprox_angle_0_not_clipped=mean_angle_0_not_clipped)
return aprox_dict