def test(self):
logging.basicConfig(level=logging.WARN)
outstring = ""
outname = "pet_output%s.tsv" % outstring
fhandle = open(outname, 'w')
num_agents = 1000
num_tsteps = 100
wealth = np.random.random_sample(num_agents)
interaction = np.random.random_sample(num_agents)
responsible = np.random.random_sample(num_agents)
time = np.random.random_sample(num_agents)
gender = np.random.randint(1, 3, num_agents)
pet = np.zeros(1000)
g = igraph.Graph()
for i in range(0, num_agents):
g.add_vertex(i)
vs = g.vs
agents = []
for i in range(0, num_agents):
traits = np.zeros((7))
traits[0] = pet[i]
traits[1] = wealth[i]
traits[2] = interaction[i]
traits[3] = responsible[i]
traits[4] = time[i]
traits[5] = gender[i]
traits[6] = 0
g.add_edge(i, vs[1])
specialagent = Person(vs[i], 7, traits)
agents.append(specialagent)
np.random.seed(347)
env = PetEnvironment(g)
time = dworp.BasicTime(num_tsteps)
scheduler = dworp.RandomOrderScheduler(np.random.RandomState(4587))
term = PetTerminator(50)
observer = dworp.ChainedObserver(PetObserver(fhandle))
sim = dworp.BasicSimulation(agents,
env,
time,
scheduler,
observer,
terminator=term)
sim.run()
fhandle.close()
with open("pet_demographics.tsv", 'w') as f:
f.write('Wealth\tInteraction\tResponsible\tTime\tGender\tPet\n')
for i in range(0, num_agents):
f.write("{}\t{}\t{}\t{}\t{}\t".format(wealth[i],
interaction[i],
responsible[i], time[i],
gender[i]))
f.write("{}\n".format(agent.state[0] for agent in agents))
def __init__(self, params, observer):
self.params = params
self.rng = np.random.RandomState(params.seed)
factory = ThingFactory(self.rng, params.similarity, params.color[0], params.color[1], params.color[2], params.color[3])
time = dworp.InfiniteTime()
scheduler = dworp.RandomOrderScheduler(self.rng)
terminator = TheTerminator()
agents = []
grid = dworp.Grid(params.grid_width, params.grid_height)
env = TheEnvironment(grid, self.rng)
for x in range(params.grid_width):
for y in range(params.grid_height):
if self.rng.uniform() < params.density:
agent = factory.create(x, y)
grid.add(agent, x, y)
agents.append(agent)
super().__init__(agents, env, time, scheduler, observer, terminator)
def test(self):
lastcountshouldbe = 4
logging.basicConfig(level=logging.WARN)
# ensuring reproducibility by setting the seed
np.random.seed(34756)
xdim = 10
ydim = 10
n_tsteps = 8000 # because we cycle through the 100 sites each time, this represents 80K events
g = igraph.Graph.Lattice([xdim, ydim],
nei=1,
directed=False,
circular=False)
agents = [Site(v) for v in g.vs]
env = AxelrodEnvironment(g)
time = dworp.BasicTime(n_tsteps)
# ensuring reproducibility by setting the seed
scheduler = dworp.RandomOrderScheduler(np.random.RandomState(4587))
observer = AxelrodObserver(1000)
term = AxelrodTerminator(1000)
sim = dworp.TwoStageSimulation(agents,
env,
time,
scheduler,
observer,
terminator=term)
sim.run()
lastcount = observer.computenumregions(0, agents, env)
print("Last Count = %d" % (lastcount))
if lastcount == lastcountshouldbe:
print("Regression test passed!")
return True
else:
print("Regression test failed! last count should be %d" %
(lastcountshouldbe))
return False
circular=False)
env = axelrod_aurora_test1.AxelrodEnvironment(g)
for k in range(0, N):
this_s_time = time.clock()
# get this simulation's seed
curseed = seedlist[place]
place = place + 1
np.random.seed(curseed)
timeobj = dworp.BasicTime(n_tsteps)
# reset all the agent states
agents = [
axelrod_aurora_test1.Site(v, num_features, num_traits)
for v in g.vs
]
# ensuring reproducibility by setting the seed
scheduler = dworp.RandomOrderScheduler(
np.random.RandomState(curseed + 1))
sim = dworp.TwoStageSimulation(agents,
env,
timeobj,
scheduler,
observer,
terminator=term)
sim.run()
lastcount = observer.computenumregions(0, agents, env)
allresults[i, k] = lastcount
this_e_time = time.clock()
alltimings[i, k] = this_e_time - this_s_time
try:
end_time = time.clock()
sim_time_minutes = float(end_time - start_time) / 60.0
print("simulation finished after %.2f minutes" % (sim_time_minutes))
# logger = logging.getLogger(__name__)
#
# def __init__(self, agents, env, time, scheduler, observer, terminator=None):
# super().__init__(agents, env, time, scheduler, observer, terminator, True)
# self.numregions = 0
logging.basicConfig(level=logging.WARN)
# ensuring reproducibility by setting the seed
np.random.seed(34756)
xdim = 10
ydim = 10
n_tsteps = 8000 # because we cycle through the 100 sites each time, this represents 80K events
g = igraph.Graph.Lattice([xdim, ydim], nei=1, directed=False, circular=False)
agents = [Site(v) for v in g.vs]
env = AxelrodEnvironment(g)
time = dworp.BasicTime(n_tsteps)
# ensuring reproducibility by setting the seed
scheduler = dworp.RandomOrderScheduler(np.random.RandomState(4587))
observer = AxelrodObserver(1000)
term = AxelrodTerminator(1000)
sim = dworp.TwoStageSimulation(agents,
env,
time,
scheduler,
observer,
terminator=term)
sim.run()
lastcount = observer.computenumregions(0, agents, env)
print("Last Count = %d" % (lastcount))
def test(self):
lastcountshouldbe = 4
logging.basicConfig(level=logging.WARN)
n_tsteps = 1000
n_tsteps = 100
n_agents = 10000
n_fps = 4
mu = np.array([0.544, 0.504, 0.466, 0.482, 0.304])
cov = np.zeros((5, 5))
cov[0, :] = [0.360000, 0.066120, 0.059520, 0.093000, 0.092040]
cov[1, :] = [0.066120, 0.336400, 0.061132, 0.061132, 0.000000]
cov[2, :] = [0.059520, 0.061132, 0.384400, 0.042284, -0.021948]
cov[3, :] = [0.093000, 0.061132, 0.042284, 0.384400, 0.098766]
cov[4, :] = [0.092040, 0.000000, -0.021948, 0.098766, 0.348100]
scenariostr = "B"
makevis = True
# n_frields: each agent has this many friends (based on the n_friends people who are geographically closest)
if scenariostr == "A":
n_friends = 2 # Scenario A
outstring = "_A"
# ensuring reproducibility by setting the seed
np.random.seed(45)
mu[0] = 0.25
elif scenariostr == "B":
n_friends = 5 # Scenario B
# ensuring reproducibility by setting the seed
np.random.seed(347)
outstring = "_B"
mu[0] = 0.50
else:
n_friends = 20 # Scenario C
outstring = "_C"
# ensuring reproducibility by setting the seed
np.random.seed(5769)
mu[0] = 0.75
n_friends = 5 # constant
personalities = np.random.multivariate_normal(mu, cov, n_agents)
personalities[personalities > 1] = 1.0
personalities[personalities < 0] = 0.0
wealth = np.random.normal(300000, 100000, n_agents)
wealth[wealth > 600000] = 600000
wealth[wealth < 10000] = 10000
offsets_lat = np.random.random((n_agents, 1))
offsets_lon = np.random.random((n_agents, 1))
lat = offsets_lon + 37.4316 # deg north
lon = offsets_lat + 78.6569 # deg west
gender = np.random.randint(0, 1, (n_agents, 1))
education = np.random.randint(0, 4, (n_agents, 1))
# colddrinks = np.random.normal(0.80, 0.15, n_agents)
colddrinks = np.random.normal(0.90, 0.1, n_agents)
colddrinks[colddrinks > 1] = 1
colddrinks[colddrinks < 0] = 0
# eatingout = np.random.normal(0.70,0.10,n_agents)
eatingout = np.random.normal(0.90, 0.10, n_agents)
eatingout[eatingout > 1] = 1
eatingout[eatingout < 0] = 0
envaware = np.random.random((n_agents, 1))
g = igraph.Graph()
for i in range(0, n_agents):
g.add_vertex(i)
vs = g.vs
agents = []
for i in range(0, n_agents):
traits = np.zeros((14))
traits[0] = 0 # initially noone uses the reusable straw
traits[1] = eatingout[i]
traits[2] = envaware[i]
traits[3:8] = personalities[i, :]
traits[8] = wealth[i]
traits[9] = lat[i]
traits[10] = lon[i]
traits[11] = gender[i]
traits[12] = education[i]
traits[13] = colddrinks[i]
difflat = lat - lat[i]
difflon = lon - lon[i]
distsq = np.power(difflat, 2) + np.power(difflon, 2)
sorted = np.argsort(distsq, axis=0)
friends = sorted[1:n_friends + 1]
for j in range(0, len(friends)):
g.add_edge(i, int(friends[j]))
curagent = Person(vs[i], 14, traits)
agents.append(curagent)
env = EEEnvironment(g)
time = dworp.BasicTime(n_tsteps)
# ensuring reproducibility by setting the seed
scheduler = dworp.RandomOrderScheduler(np.random.RandomState(4587))
outname = "outputs%s.tsv" % (outstring)
fhandle = open(outname, 'w')
myobserver = EEObserver(fhandle)
#vis_flag = args.vis and 'pygame' in sys.modules
vis_flag = makevis and 'pygame' in sys.modules
if vis_flag:
print("vis_flag is True")
else:
print("vis_flag is False")
# create and run one realization of the simulation
observer = dworp.ChainedObserver(myobserver, )
if vis_flag:
observer.append(dworp.PauseAtEndObserver(3))
pgr = PyGameRenderer(1, n_fps, n_tsteps + 1)
observer.append(pgr)
#with open("eatingout.tsv",'w') as f:
# for i in range(0,n_agents):
# f.write('%f\t%f' % (lat[i],lon[i]))
# f.write('\t%f\n' % (eatingout[i]))
# f.close()
initeatingout = eatingout[0:]
term = EETerminator(100)
sim = dworp.BasicSimulation(agents,
env,
time,
scheduler,
observer,
terminator=term)
sim.run()
fhandle.close()
# eatingout, age
age = np.random.randint(16, 65, 1000)
with open("demog.tsv", 'w') as f:
for i in range(0, n_agents):
f.write('%f\t%f' % (lat[i], lon[i]))
f.write('\t%f\t%f' % (wealth[i], gender[i]))
f.write(
'\t%d\t%d' %
(agents[i].state[0], agents[i].state[agents[i].past_i]))
f.write('\t%f\t%d\t%f\t%f\t%f\t%f\t%f\t%f\n' %
(initeatingout[i], age[i], colddrinks[i],
personalities[i, 0], personalities[i, 1],
personalities[i, 2], personalities[i, 3],
personalities[i, 4]))
f.close()
if vis_flag:
filename_list = pgr.filename_list
seconds_per_frame = 1.0 / n_fps
frame_delay = str(int(seconds_per_frame * 100))
#command_list = ['convert', '-delay', frame_delay, '-loop', '0'] + filename_list + ['anim.gif']
command_list = ['convert', '-delay', frame_delay, '-loop', '0'
] + filename_list + ['anim%s.gif' % (outstring)]
pdb.set_trace()
try:
# Use the "convert" command (part of ImageMagick) to build the animation
subprocess.call(command_list)
except:
print("couldnt create the animation")
pass
# Earlier, we saved an image file for each frame of the animation. Now
# that the animation is assembled, we can (optionally) delete those files
for filename in filename_list:
os.remove(filename)
return
lastcount = myobserver.computenumreusablestrawusers(0, agents, env)
print("Last Count = %d" % (lastcount))
if lastcount == lastcountshouldbe:
print("Regression test passed!")
return True
else:
print("Regression test failed! last count should be %d" %
(lastcountshouldbe))
return False
def test(self):
logging.basicConfig(level=logging.WARN)
outstring = "_B"
outname = "outputs%s.tsv" % (outstring)
fhandle = open(outname, 'w')
num_agents = 5000
num_tsteps = 150
wealth = np.random.normal(100000, 20000, num_agents)
offsets_lat = np.random.random((num_agents, 1))
offsets_lon = np.random.random((num_agents, 1))
lat = offsets_lon + 37.4316 # deg north
lon = offsets_lat + 78.6569 # deg west
SA = np.random.random((num_agents, 1))
phone = np.random.random_integers(0, 2, num_agents)
g = igraph.Graph()
for i in range(0, num_agents):
g.add_vertex(i)
vs = g.vs
agents = []
for i in range(0, num_agents):
traits = np.zeros((5))
traits[0] = phone[i]
traits[1] = wealth[i]
traits[2] = lat[i]
traits[3] = lon[i]
traits[4] = SA[i]
difflat = lat - lat[i]
difflon = lon - lon[i]
distsq = np.power(difflat, 2) + np.power(difflon, 2)
sorted = np.argsort(distsq, axis=0)
n_friends = 5
friends = sorted[1: n_friends+1]
for j in range(0,len(friends)):
g.add_edge(i,int(friends[j]))
curagent = Person(vs[i], 5, traits)
agents.append(curagent)
np.random.seed(347)
makevis = True
vis_flag = makevis and 'pygame' in sys.modules
if vis_flag:
print("vis_flag is True")
else:
print("vis_flag is False")
# vis does not support different colors
# color
#
# s = ["blue", "orange"]
# params = SegregationParams(density, similarity, grid_size, seed, colors)
# create and run one realization of the simulation
observer = dworp.ChainedObserver(PhoneObserver(fhandle))
n_fps = 4
if vis_flag:
observer.append(dworp.PauseAtEndObserver(3))
pgr = PyGameRenderer(1, n_fps, num_tsteps + 1)
observer.append(pgr)
env = PhoneEnvironment(g)
time = dworp.BasicTime(num_tsteps)
scheduler = dworp.RandomOrderScheduler(np.random.RandomState(4587))
term = PhoneTerminator(50)
sim = dworp.BasicSimulation(agents, env, time, scheduler, observer, terminator=term)
sim.run()
fhandle.close()
with open("demographics.tsv",'w') as f:
f.write('Lat\tLon\tWealth\tSA\tPhone\n')
for i in range(0,num_agents):
f.write('{}\t{}\t'.format(lat[i], lon[i]))
f.write('{}\t'.format(wealth[i]))
f.write('{}'.format(SA[i]))
if agents[i].state[0] == 0:
f.write('\tNone\n')
elif agents[i].state[0] == 1:
f.write('\tApple\n')
elif agents[i].state[0] == 2:
f.write('\tAndroid\n')
f.close()
