+import distributions
+import matplotlib.pyplot as plt
+import streamlit as st
+
+
+[docs]def show_points(
+
s_polygon,
+
samples,
+
radius,
+
circle_segment_radii=None,
+
center_point=None,
+
obstacles=None,
+
):
+
if obstacles is None:
+
obstacles = []
+
box = distributions.__get_bounding_box(s_polygon)
+
exterior = list(s_polygon.exterior.coords)
+
fig = plt.figure()
+
ax = fig.add_subplot(1, 1, 1)
+
for elem in samples:
+
ax.add_patch(plt.Circle(radius=radius / 2, xy=elem, fill=False))
+
ax.add_patch(plt.Circle(radius=0.1, xy=elem, color="r"))
+
+
if circle_segment_radii is not None:
+
for circle_segment in circle_segment_radii:
+
ax.add_patch(
+
plt.Circle(
+
radius=circle_segment[0], xy=center_point, fill=False
+
)
+
)
+
ax.add_patch(
+
plt.Circle(
+
radius=circle_segment[1], xy=center_point, fill=False
+
)
+
)
+
+
n = len(exterior)
+
i = 0
+
while True:
+
following = (i + 1) % n
+
x_value = [exterior[i][0], exterior[following][0]]
+
y_value = [exterior[i][1], exterior[following][1]]
+
plt.plot(x_value, y_value, color="blue")
+
+
i += 1
+
if following == 0:
+
break
+
+
for obstacle in obstacles:
+
n = len(obstacle)
+
i = 0
+
while True:
+
following = (i + 1) % n
+
x_value = [obstacle[i][0], obstacle[following][0]]
+
y_value = [obstacle[i][1], obstacle[following][1]]
+
plt.plot(x_value, y_value, color="black")
+
+
i += 1
+
if following == 0:
+
break
+
+
plt.xlim(box[0][0], box[1][0])
+
plt.ylim(box[0][1], box[1][1])
+
plt.axis("equal")
+
st.pyplot(fig)
+
+
+[docs]def main():
+
default_polygon = [(0.0, 0.0), (10.0, 0.0), (10.0, 10.0), (0.0, 10.0)]
+
agents = density = 0
+
+
with st.sidebar:
+
st.text("General settings")
+
distance_to_agents = st.slider(
+
"Distance between the center points of the agents",
+
0.1,
+
2.0,
+
value=0.3,
+
)
+
distance_to_polygon = st.slider(
+
"Distance between agents and walls", 0.1, 2.0, value=0.3
+
)
+
st.text("Polygon settings")
+
corners = st.number_input(
+
"Number of edges in polygon", 3, value=len(default_polygon)
+
)
+
col1, col2 = st.columns(2)
+
x_values, y_values = [], []
+
with col1:
+
for i in range(corners):
+
x_values.append(
+
st.number_input(
+
f"X{i + 1}",
+
value=default_polygon[i][0]
+
if i < len(default_polygon)
+
else 0.0,
+
step=1.0,
+
)
+
)
+
with col2:
+
for i in range(corners):
+
y_values.append(
+
st.number_input(
+
f"Y{i + 1}",
+
value=default_polygon[i][1]
+
if i < len(default_polygon)
+
else 0.0,
+
step=1.0,
+
)
+
)
+
seed = st.number_input(
+
"Set a seed for random number generation.\n"
+
"Agent Distribution with the same settings will lead to the same output.\n"
+
"If the seed is set to zero the distribution will be random each time",
+
0,
+
)
+
if seed == 0:
+
seed = None
+
st.text(
+
"Hole settings:\n"
+
"Holes define Spaces where no Agents must be placed."
+
)
+
obstacle_count = st.number_input("Number of Holes", 0, value=0)
+
obstacle_corners = []
+
obstacle_values = []
+
for i in range(obstacle_count):
+
obstacle_corners.append(
+
st.number_input(f"Number of edges in hole {i + 1}", 3, value=4)
+
)
+
col3, col4 = st.columns(2)
+
with col3:
+
for i in range(obstacle_count):
+
obstacle_values.append([])
+
st.text(f"Settings for")
+
# the rest of the text can be found in column 4
+
for j in range(obstacle_corners[i]):
+
obstacle_values[i].append(
+
[
+
st.number_input(
+
f"X{j + 1} in hole {i + 1}",
+
value=0.0,
+
step=1.0,
+
)
+
]
+
)
+
with col4:
+
for i in range(obstacle_count):
+
st.text(f"hole {i + 1}")
+
# this text is placed so the text "Settings for hole <i+1> aligns through columns 3 & 4
+
for j in range(obstacle_corners[i]):
+
obstacle_values[i][j].append(
+
st.number_input(
+
f"Y{j + 1} in hole {i + 1}", value=0.0, step=1.0
+
)
+
)
+
+
obstacles = []
+
for i in range(obstacle_count):
+
obstacles.append([])
+
for j in range(obstacle_corners[i]):
+
obstacles[i].append(
+
(obstacle_values[i][j][0], obstacle_values[i][j][1])
+
)
+
polygon = []
+
for x, y in zip(x_values, y_values):
+
polygon.append((x, y))
+
s_polygon = distributions.shply.Polygon(polygon, obstacles)
+
area = s_polygon.area
+
distribution_type = st.radio(
+
"How to distribute agents?",
+
(
+
"place by number/density",
+
"place in circles",
+
"place till full",
+
"place by percentage",
+
),
+
)
+
+
if distribution_type == "place in circles":
+
st.text(
+
"Instructions: "
+
"Agents will be distributed inside both the polygon and circle segments.\n"
+
"Agents do not keep distance to circle segment borders like polygons.\n"
+
"You need to select a center point for all circle segments.\n"
+
"You also need to select the range of each circle segment.\n"
+
"Circle segments musst not overlap or have negativ values.\n"
+
"General settings can be made in the left sidebar."
+
)
+
style = st.radio(
+
f"How to choose number of agents distributed inside circle segments?",
+
("density", "number"),
+
)
+
circle_count = st.number_input("number of circles", 1)
+
col1, col2 = st.columns(2)
+
min_values, max_values = [], []
+
densities, agents = [], []
+
default_center = (5.0, 5.0)
+
with col1:
+
mid_x = st.number_input(
+
"Center X", value=default_center[0], step=1.0
+
)
+
for i in range(circle_count):
+
min_values.append(
+
st.number_input(
+
f"inner radius for Circle segment {i + 1}", step=1.0
+
)
+
)
+
with col2:
+
mid_y = st.number_input(
+
"Center Y", value=default_center[1], step=1.0
+
)
+
for i in range(circle_count):
+
max_values.append(
+
st.number_input(
+
f"outer radius for Circle segment {i + 1}", step=1.0
+
)
+
)
+
+
center_point = (mid_x, mid_y)
+
circle_segment_radii = []
+
for i in range(circle_count):
+
circle_segment_radii.append((min_values[i], max_values[i]))
+
+
for i in range(circle_count):
+
if style == "density":
+
densities.append(
+
st.slider(
+
f"Agents / m² in Circle segment {i + 1}",
+
0.1,
+
round(
+
(
+
69.2636
+
* (distance_to_agents ** (-1.89412))
+
/ 100
+
),
+
2,
+
),
+
key=i,
+
)
+
)
+
# the function for the maximum selectable density is an estimate of the real maximum density possible
+
# the estimation was done by Power-Regression
+
elif style == "number":
+
temp_min, temp_max = (
+
circle_segment_radii[i][0],
+
circle_segment_radii[i][1],
+
)
+
area_small = distributions.__intersecting_area_polygon_circle(
+
center_point, temp_min, s_polygon
+
)
+
area_big = distributions.__intersecting_area_polygon_circle(
+
center_point, temp_max, s_polygon
+
)
+
area_segment = area_big - area_small
+
agents.append(
+
st.slider(
+
f"Agents in Circle segment {i + 1}",
+
1,
+
round(area_segment * 8),
+
key=i,
+
)
+
)
+
+
button_clicked = st.button("distribute agents")
+
if button_clicked:
+
samples = []
+
if style == "density":
+
samples = distributions.distribute_in_circles_by_density(
+
polygon=s_polygon,
+
distance_to_agents=distance_to_agents,
+
distance_to_polygon=distance_to_polygon,
+
center_point=center_point,
+
circle_segment_radii=circle_segment_radii,
+
densities=densities,
+
seed=seed,
+
)
+
elif style == "number":
+
samples = distributions.distribute_in_circles_by_number(
+
polygon=s_polygon,
+
distance_to_agents=distance_to_agents,
+
distance_to_polygon=distance_to_polygon,
+
center_point=center_point,
+
circle_segment_radii=circle_segment_radii,
+
numbers_of_agents=agents,
+
seed=seed,
+
)
+
st.text(f"Below should be a plot containing {len(samples)} agents")
+
show_points(
+
s_polygon,
+
samples,
+
distance_to_agents,
+
circle_segment_radii,
+
center_point,
+
obstacles,
+
)
+
else:
+
show_points(
+
s_polygon,
+
[],
+
distance_to_agents,
+
circle_segment_radii,
+
center_point,
+
obstacles,
+
)
+
+
if distribution_type == "place by number/density":
+
st.text(
+
"Instructions: "
+
"a certain amount of agents will be distributed inside the polygon.\n"
+
"You can enter this amount either by a specific number or a density.\n"
+
"General settings can be made in the left sidebar."
+
)
+
style = st.radio(
+
"How to choose number of agents?", ("density", "number")
+
)
+
if style == "density":
+
density = st.slider(
+
"Agents / m²",
+
0.1,
+
round((69.2636 * (distance_to_agents ** (-1.89412)) / 100), 2),
+
)
+
# the function for the maximum selectable density is an estimate of the real maximum density possible
+
# the estimation was done by Power-Regression
+
elif style == "number":
+
agents = st.slider("Agents", 1, round(area * 8))
+
+
button_clicked = st.button("distribute agents")
+
+
if button_clicked:
+
samples = []
+
if style == "density":
+
samples = distributions.distribute_by_density(
+
polygon=s_polygon,
+
density=density,
+
distance_to_agents=distance_to_agents,
+
distance_to_polygon=distance_to_polygon,
+
seed=seed,
+
)
+
elif style == "number":
+
samples = distributions.distribute_by_number(
+
polygon=s_polygon,
+
number_of_agents=agents,
+
distance_to_agents=distance_to_agents,
+
distance_to_polygon=distance_to_polygon,
+
seed=seed,
+
)
+
st.text(f"Below should be a plot containing {len(samples)} agents")
+
show_points(
+
s_polygon=s_polygon,
+
samples=samples,
+
radius=distance_to_agents,
+
obstacles=obstacles,
+
)
+
else:
+
show_points(
+
s_polygon=s_polygon,
+
samples=[],
+
radius=distance_to_agents,
+
obstacles=obstacles,
+
)
+
+
if distribution_type == "place till full":
+
st.text(
+
"Instructions: "
+
"agents will be distributed all over the polygon.\n"
+
"You can enter how many tries there should be to "
+
"find another place to put an agent before a polygon is considered filled.\n"
+
"General settings can be made in the left sidebar"
+
)
+
max_iterations = st.number_input(
+
"maximum tries to find the first point", 1, value=10_000, step=100
+
)
+
k = st.number_input(
+
"maximum tries to find a point around a reference point",
+
1,
+
value=30,
+
step=1,
+
)
+
button_clicked = st.button("distribute agents")
+
+
if button_clicked:
+
samples = distributions.distribute_till_full(
+
polygon=s_polygon,
+
distance_to_agents=distance_to_agents,
+
distance_to_polygon=distance_to_polygon,
+
seed=seed,
+
max_iterations=max_iterations,
+
k=k,
+
)
+
st.text(f"Below should be a plot containing {len(samples)} agents")
+
show_points(
+
s_polygon=s_polygon,
+
samples=samples,
+
radius=distance_to_agents,
+
obstacles=obstacles,
+
)
+
else:
+
show_points(
+
s_polygon=s_polygon,
+
samples=[],
+
radius=distance_to_agents,
+
obstacles=obstacles,
+
)
+
+
if distribution_type == "place by percentage":
+
st.text(
+
"Instructions: "
+
"distributes agents inside the polygon.\n"
+
"The polygon will be filled to a certain percentage.\n"
+
"You can also enter how many tries there should be to "
+
"find another place to put an agent before a polygon is considered filled.\n"
+
"General settings can be made in the left sidebar"
+
)
+
percent = st.slider("Percent filled: ", 1, 100)
+
max_iterations = st.number_input(
+
"maximum tries to find the first point", 1, value=10_000, step=100
+
)
+
k = st.number_input(
+
"maximum tries to find a point around a reference point",
+
1,
+
value=30,
+
step=1,
+
)
+
button_clicked = st.button("distribute agents")
+
+
if button_clicked:
+
samples = distributions.distribute_by_percentage(
+
polygon=s_polygon,
+
percent=percent,
+
distance_to_agents=distance_to_agents,
+
distance_to_polygon=distance_to_polygon,
+
seed=seed,
+
max_iterations=max_iterations,
+
k=k,
+
)
+
st.text(f"Below should be a plot containing {len(samples)} agents")
+
show_points(
+
s_polygon=s_polygon,
+
samples=samples,
+
radius=distance_to_agents,
+
obstacles=obstacles,
+
)
+
else:
+
show_points(
+
s_polygon=s_polygon,
+
samples=[],
+
radius=distance_to_agents,
+
obstacles=obstacles,
+
)
+
+
+if __name__ == "__main__":
+ main()
+
+
+
+
+
+
+
+
+
+
+
+
+