def test_default_plots_True_plotcallback(self):
agent = agents.PpoAgent("CartPole-v0")
p = plot.Loss()
r = plot.Rewards()
c = agent._add_plot_callbacks([r], True, [p])
assert p in c
assert r in c
def test_default_plots_None_plotcallback(self):
agent = agents.PpoAgent("CartPole-v0")
p = plot.Loss()
r = plot.Rewards()
c = agent._prepare_callbacks([r], None, [p])
assert not p in c
assert r in c
def test_prepare_callbacks(self):
agent = agents.PpoAgent("CartPole-v0")
c = [plot.ToMovie(), plot.Rewards()]
d = agent._prepare_callbacks(c, default_plots=None, default_plot_callbacks=[])
assert isinstance(d[0], plot._PreProcess)
assert isinstance(d[1], plot.Rewards)
assert isinstance(d[-2], plot._PostProcess)
assert isinstance(d[-1], plot.ToMovie)
def test_train_single_episode(self):
for backend in get_backends(PpoAgent):
ppo = agents.PpoAgent(gym_env_name=_env_name, backend=backend)
count = log._CallbackCounts()
ppo.train([log.Agent(), count, duration._SingleEpisode()])
assert count.gym_init_begin_count == count.gym_init_end_count == 1
assert count.gym_step_begin_count == count.gym_step_end_count
assert count.gym_step_begin_count < 10 + count.gym_reset_begin_count
def test_save_load(self):
for backend in get_backends(PpoAgent):
ppo = agents.PpoAgent(gym_env_name=_step_count_name, backend=backend)
ppo.train([duration._SingleEpisode()], default_plots=False)
temp_dir = ppo.save()
ppo = agents.load(temp_dir)
ppo.play(default_plots=False, num_episodes=1, callbacks=[])
easyagents.backends.core._rmpath(temp_dir)
def test_play_single_episode(self):
for backend in get_backends(PpoAgent):
ppo = agents.PpoAgent(gym_env_name=_env_name, backend=backend)
count = log._CallbackCounts()
cb = [log.Agent(), count, duration._SingleEpisode()]
ppo.train(duration._SingleEpisode())
ppo.play(cb)
assert count.gym_init_begin_count == count.gym_init_end_count == 1
assert count.gym_step_begin_count == count.gym_step_end_count <= 10
def test_train_multiple_subplots(self):
agent = agents.PpoAgent("CartPole-v0")
agent.train([
duration._SingleIteration(),
plot.State(),
plot.Rewards(),
plot.Loss(),
plot.Steps()
])
def test_every(self):
agent = agents.PpoAgent(_line_world_name)
every = save.Every(num_evals_between_save=1)
agent.train([every],
num_iterations_between_eval=1,
num_episodes_per_iteration=10,
num_iterations=3,
default_plots=False)
os.path.isdir(every.directory)
assert len(every.saved_agents) == 4
for (episode, reward, dir) in every.saved_agents:
os.path.isdir(dir)
def test_train_tomovie_with_filename(self):
f = tempfile.NamedTemporaryFile(delete=False)
filepath = f.name
f.close()
os.remove(filepath)
assert not os.path.isfile(filepath)
filepath += '.gif'
agent = agents.PpoAgent("CartPole-v0")
m = plot.ToMovie(filepath=filepath, fps=10)
agent.train([duration._SingleIteration(), plot.Rewards(), m])
try:
os.remove(m.filepath)
except:
pass
def test_best(self):
agent = agents.PpoAgent(_line_world_name)
best = save.Best()
agent.train([best],
num_iterations_between_eval=1,
num_episodes_per_iteration=10,
num_iterations=3,
default_plots=False)
os.path.isdir(best.directory)
assert len(best.saved_agents) > 0
(episode, reward, dir) = best.saved_agents[0]
os.path.isdir(dir)
agent2 = easyagents.agents.load(dir)
assert agent2
agent2.evaluate()
def test_train_plotsteps(self):
agent = agents.PpoAgent("CartPole-v0")
agent.train([duration._SingleIteration(), plot.Steps()])
def test_train_plotsteprewards(self):
ppoAgent = agents.PpoAgent('CartPole-v0')
ppoAgent.train([plot.StepRewards(), duration.Fast()])
def test_single_episode(self):
agent = agents.PpoAgent("CartPole-v0")
count=log._CallbackCounts()
agent.train([duration._SingleEpisode(), log._Callbacks(), count])
assert count.train_iteration_begin_count == 1
def test_default_plots_None_nocallback(self):
agent = agents.PpoAgent("CartPole-v0")
p = plot.Loss()
c = agent._add_plot_callbacks([], None, [p])
assert p in c
def test_play_plotrewards(self):
agent = agents.PpoAgent("CartPole-v0")
agent.train([duration._SingleIteration()])
agent.play([plot.Rewards()])
def test_play_plotstate(self):
agent = agents.PpoAgent("CartPole-v0")
agent.train([duration._SingleEpisode()])
agent.play([plot.State()])
def test_cartpole_log_iteration(self):
ppo = agents.PpoAgent(gym_env_name="CartPole-v0", backend='tfagents')
ppo.train([log.Iteration(), duration._SingleIteration()])
def test_log_agent(self):
agent = agents.PpoAgent(_step_count_name)
agent.train([log.Agent(), duration.Fast()])
def test_default_plots_None_durationcallback(self):
agent = agents.PpoAgent("CartPole-v0")
p = plot.Loss()
c = agent._prepare_callbacks([duration.Fast()], None, [p])
assert p in c
def test_default_plots_False_nocallback(self):
agent = agents.PpoAgent("CartPole-v0")
p = plot.Loss()
c = agent._prepare_callbacks([], False, [p])
assert not p in c
def test_log_step(self):
agent = agents.PpoAgent(_step_count_name)
agent.train([log.Step(), duration._SingleIteration()])
def test_train_tomovie(self):
agent = agents.PpoAgent("CartPole-v0")
agent.train(
[duration._SingleIteration(),
plot.Rewards(),
plot.ToMovie()])
def test_log_callbacks(self):
agent = agents.PpoAgent(_step_count_name)
agent.train([log._Callbacks(), duration._SingleIteration()])
def test_log_iteration(self):
agent = agents.PpoAgent(env_name)
agent.train([log.Iteration(), duration._SingleIteration()])
