def plot_net_state_actions(self, net):
"""
plot all state actions
"""
fig, figure = plt.subplots()
figure.set_xlim([-.5, self.w - 1 +.5])
figure.set_ylim([-.5, self.h - 1 + .5])
width_range = np.arange(self.w)
height_range = np.arange(self.h)
plt.xticks(width_range)
plt.yticks(height_range)
figure.set_yticks((height_range[:-1] + 0.5), minor=True)
figure.set_xticks((width_range[:-1] + 0.5), minor=True)
figure.grid(which='minor', axis='both', linestyle='-')
plt.scatter([7],[7], c= 'green',s=300) # reward
plt.scatter([4],[2], c= 'red',s=300) # sink
# [-1] = stay, [0] = up, [1] = right, [2] = down, [3] = left
state_actions = [[], [], [], [], []]
markers = ['^', '>', 'v', '<', 'o']
colors = ['c', 'm', 'k', 'y', 'b']
for i in range(self.w):
for j in range(self.h):
state = State(i, j)
state = np.zeros([1,2])+state.toArray()
next_action = int(net.predict(state))
state_actions[next_action].append((i, j))
for direction, marker, color in zip(state_actions, markers, colors):
xs, ys = [], []
for x, y in direction:
xs.append(x)
ys.append(y)
figure.scatter(xs, ys, marker=marker, s=50, color=color)
plt.show()
