Created
March 13, 2020 10:52
-
-
Save Quasimondo/3a8f2d0d02f708b348b7d0504c8de41c to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| import os | |
| import numpy as np | |
| import threading | |
| from imageio import imwrite,imread | |
| from scipy.ndimage import zoom | |
| class Corona(): | |
| def __init__(self): | |
| self.infectionProbability = 0.02 | |
| self.lethalProbability = 0.02 | |
| self.stepsUntilContagious_min = 10 | |
| self.stepsUntilContagious_max = 30 | |
| self.stepsUntilSymptoms_min = 50 | |
| self.stepsUntilSymptoms_max = 80 | |
| self.stepsUntilHealed_min = 200 | |
| self.stepsUntilHealed_max = 300 | |
| self.stepsUntilReceptive_min = 500 | |
| self.stepsUntilReceptive_max = 700 | |
| self.contaminationRadius = 1 | |
| self.symptoms_show = 0.5 | |
| self.travel_distance = 1 | |
| self.travel_probability = 0.2 | |
| self.air_travel_probability = 0.0 | |
| self.social_distance_probability = 0.0 | |
| #self.no_symptoms_avoid_others = 0.1 | |
| #self.symptoms_avoid_others = 0.75 | |
| #self.immune_avoid_others = 0.01 | |
| class World(): | |
| def __init__(self,corona=None,world_width=1024,world_height=1024,agents=1000,initialInfections=10): | |
| self.states = ["HEALTHY_NOT_IMMUNE","INFECTED_NO_SYMPTOMS","INFECTED_SYMPTOMS","HEALTHY_IMMUNE","DEAD"] | |
| self.colors = [[86,180,233],[240,228,66],[213,94,0],[0,114,178],[204,121,167]] | |
| self.emojis = [] | |
| self.emojis.append(imread("emojis/nonimmune.png",pilmode="RGB")) | |
| self.emojis.append(imread("emojis/contagious.png",pilmode="RGB")) | |
| self.emojis.append(imread("emojis/symptoms.png",pilmode="RGB")) | |
| self.emojis.append(imread("emojis/immune.png",pilmode="RGB")) | |
| self.emojis.append(imread("emojis/dead.png",pilmode="RGB")) | |
| self.emojiScale = self.emojis[0].shape[0] | |
| self.world = np.zeros((world_height,world_width),dtype=int)-1 | |
| self.corona = corona | |
| self.agents = [] | |
| startingPositions = np.random.choice(np.arange(world_width*world_height),size=16,replace=False) | |
| infected = np.random.choice(np.arange(agents),size=initialInfections,replace=False) | |
| for i in range(16): | |
| a = Agent(startingPositions[i]%world_width,int(startingPositions[i]/world_width),1 if i in infected else 0,will_show_symptoms=np.random.random()<corona.symptoms_show,will_die_if_infected=np.random.random()<corona.lethalProbability) | |
| self.world[a.y,a.x] = len(self.agents) | |
| self.agents.append(a) | |
| for i in range(16,agents): | |
| found = False | |
| while not found: | |
| a = self.agents[np.random.randint(i)] | |
| x = a.x + np.random.randint(-2,3) | |
| y = a.y + np.random.randint(-2,3) | |
| if x>=0 and x < self.world.shape[1] and y>=0 and y < self.world.shape[0] and self.world[y,x]==-1: | |
| found =True | |
| a = Agent(x,y,1 if i in infected else 0,will_show_symptoms=np.random.random()<corona.symptoms_show,will_die_if_infected=np.random.random()<corona.lethalProbability) | |
| self.world[a.y,a.x] = len(self.agents) | |
| self.agents.append(a) | |
| def update(self): | |
| order = np.random.permutation(len(self.agents)) | |
| for i in order: | |
| self.agents[i].update(self.world,self.corona,self.agents) | |
| def render(self): | |
| t = np.zeros((self.world.shape[0]*self.emojiScale,self.world.shape[1]*self.emojiScale,3),dtype=np.uint8)+255 | |
| for agent in self.agents: | |
| t[agent.y*self.emojiScale:agent.y*self.emojiScale+self.emojiScale,agent.x*self.emojiScale:agent.x*self.emojiScale+self.emojiScale] = self.emojis[agent.state] | |
| return t | |
| class Agent(): | |
| def __init__(self,x,y,state,will_show_symptoms,will_die_if_infected): | |
| self.state = state | |
| self.infection_time = -1 if state==0 else 0 | |
| self.x = x | |
| self.y = y | |
| self.will_show_symptoms = will_show_symptoms | |
| self.will_die = will_die_if_infected | |
| self.target_x = x | |
| self.target_y = y | |
| self.is_contagious = False | |
| def update(self,world,corona,agents): | |
| if self.state==4: | |
| return | |
| if self.infection_time>-1: | |
| self.infection_time+=1 | |
| if (self.x == self.target_x and self.y == self.target_y) or np.random.random()<0.2: | |
| x_from = max(0,self.x-corona.travel_distance) | |
| x_to = min(world.shape[1],self.x+corona.travel_distance+1) | |
| y_from = max(0,self.y-corona.travel_distance) | |
| y_to = min(world.shape[0],self.y+corona.travel_distance+1) | |
| self.target_x = np.random.randint(x_from,x_to) | |
| self.target_y = np.random.randint(y_from,y_to) | |
| if self.state == 1: | |
| if self.will_show_symptoms: | |
| th = np.random.randint(corona.stepsUntilSymptoms_min,corona.stepsUntilSymptoms_max) | |
| if self.infection_time > th: | |
| self.state = 2 | |
| th = np.random.randint(corona.stepsUntilHealed_min,corona.stepsUntilHealed_max) | |
| if self.infection_time > th: | |
| self.state = 3 | |
| self.is_contagious = False | |
| elif self.state==2: | |
| if self.will_die and np.random.random()<0.1: | |
| self.state=4 | |
| self.infection_time = -1 | |
| self.is_contagious = False | |
| return | |
| else: | |
| th = np.random.randint(corona.stepsUntilHealed_min,corona.stepsUntilHealed_max) | |
| if self.infection_time > th: | |
| self.is_contagious = False | |
| if self.will_die: | |
| self.infection_time = -1 | |
| self.state=4 | |
| return | |
| else: | |
| self.state = 3 | |
| elif self.state==3: | |
| th = np.random.randint(corona.stepsUntilReceptive_min,corona.stepsUntilReceptive_max) | |
| if self.infection_time > th: | |
| self.state = 0 | |
| if not self.is_contagious and (self.state==1 or self.state==2): | |
| th = np.random.randint(corona.stepsUntilContagious_min,corona.stepsUntilContagious_max) | |
| #print("check contagious",self.infection_time,th) | |
| if self.infection_time > th: | |
| self.is_contagious = True | |
| x_from = max(0,self.x-corona.contaminationRadius) | |
| x_to = min(world.shape[1],self.x+corona.contaminationRadius+1) | |
| y_from = max(0,self.y-corona.contaminationRadius) | |
| y_to = min(world.shape[0],self.y+corona.contaminationRadius+1) | |
| if self.state == 0: | |
| for y in range(y_from,y_to): | |
| if self.state > 0: | |
| break | |
| for x in range(x_from,x_to): | |
| if x==self.x and y==self.y: | |
| continue | |
| if world[y,x]>-1 and agents[world[y,x]].is_contagious: | |
| #print("check infection") | |
| if np.random.random()<corona.infectionProbability: | |
| #print("infected!") | |
| self.state = 1 | |
| self.infection_time = 0 | |
| break | |
| if self.state != 2 and np.random.random() < corona.travel_probability: | |
| if np.random.random()<corona.air_travel_probability: | |
| empty = np.where(world==-1) | |
| idx = np.random.randint(len(empty[0])) | |
| px = empty[1][idx] | |
| py = empty[0][idx] | |
| world[py,px] = world[self.y,self.x] | |
| world[self.y,self.x] = -1 | |
| self.x = px | |
| self.y = py | |
| else: | |
| x_from = max(0,self.x-1) | |
| x_to = min(world.shape[1],self.x+2) | |
| y_from = max(0,self.y-1) | |
| y_to = min(world.shape[0],self.y+2) | |
| bestOption=None | |
| if np.random.random()<corona.social_distance_probability: | |
| bestScore = -1 | |
| for y in range(y_from,y_to): | |
| for x in range(x_from,x_to): | |
| if world[y,x]==-1 or (x==self.x and y==self.y): | |
| score = np.sum(world[y-1:y+2,x-1:x+2]==-1) | |
| if score>bestScore: | |
| bestScore = score | |
| bestOption = [[x,y]] | |
| elif score==bestScore: | |
| bestOption.append([x,y]) | |
| else: | |
| bestScore = 1000000000 | |
| for y in range(y_from,y_to): | |
| for x in range(x_from,x_to): | |
| if world[y,x]==-1 or (x==self.x and y==self.y): | |
| dx = self.target_x - x | |
| dy = self.target_y - y | |
| if dx*dx+dy*dy<bestScore: | |
| bestScore = dx*dx+dy*dy | |
| bestOption = [[x,y]] | |
| elif dx*dx+dy*dy==bestScore: | |
| bestOption.append([x,y]) | |
| if len(bestOption)==1: | |
| bo = bestOption[0] | |
| else: | |
| bo = bestOption[np.random.randint(len(bestOption))] | |
| if bo[1]!=self.y or bo[0]!=self.x: | |
| world[bo[1],bo[0]] = world[self.y,self.x] | |
| world[self.y,self.x] = -1 | |
| self.x = bo[0] | |
| self.y = bo[1] | |
| def run(rounds=3000,world_width = 32,world_height = 32): | |
| exportFolder = "corona_export_01/" | |
| if not os.path.exists(exportFolder): | |
| os.mkdir(exportFolder) | |
| fileIndex = 0 | |
| corona = Corona() | |
| world = World(corona=corona,world_width=world_width,world_height=world_height,agents=700,initialInfections=1) | |
| for i in range(rounds): | |
| print(i,"/",rounds) | |
| world.update() | |
| tex = world.render() | |
| imwrite(exportFolder+str(fileIndex).zfill(5)+".jpg",zoom(tex,(0.75,0.75,1),order=0)) | |
| fileIndex += 1 | |
| run() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment