mirror of
https://github.com/Farama-Foundation/Gymnasium.git
synced 2025-09-01 02:17:19 +00:00
58e6aa95e5
* Make environments seedable * Fix monitor bugs - Set monitor_id before setting the infix. This was a bug that would yield incorrect results with multiple monitors. - Remove extra pid from stats recorder filename. This should be purely cosmetic. * Start uploading seeds in episode_batch * Fix _bigint_from_bytes for python3 * Set seed explicitly in random_agent * Pass through seed argument * Also pass through random state to spaces * Pass random state into the observation/action spaces * Make all _seed methods return the list of used seeds * Switch over to np.random where possible * Start hashing seeds, and also seed doom engine * Fixup seeding determinism in many cases * Seed before loading the ROM * Make seeding more Python3 friendly * Make the MuJoCo skipping a bit more forgiving * Remove debugging PDB calls * Make setInt argument into raw bytes * Validate and upload seeds * Skip box2d * Make seeds smaller, and change representation of seeds in upload * Handle long seeds * Fix RandomAgent example to be deterministic * Handle integer types correctly in Python2 and Python3 * Try caching pip * Try adding swap * Add df and free calls * Bump swap * Bump swap size * Try setting overcommit * Try other sysctls * Try fixing overcommit * Try just setting overcommit_memory=1 * Add explanatory comment * Add what's new section to readme * BUG: Mark ElevatorAction-ram-v0 as non-deterministic for now * Document seed * Move nondetermistic check into spec
73 lines
2.3 KiB
Python
73 lines
2.3 KiB
Python
import json
|
|
import os
|
|
import time
|
|
|
|
from gym import error
|
|
from gym.utils import atomic_write
|
|
|
|
class StatsRecorder(object):
|
|
def __init__(self, directory, file_prefix):
|
|
self.initial_reset_timestamp = None
|
|
self.directory = directory
|
|
self.file_prefix = file_prefix
|
|
self.episode_lengths = []
|
|
self.episode_rewards = []
|
|
self.timestamps = []
|
|
self.steps = None
|
|
self.rewards = None
|
|
|
|
self.done = None
|
|
self.closed = False
|
|
|
|
filename = '{}.stats.json'.format(self.file_prefix)
|
|
self.path = os.path.join(self.directory, filename)
|
|
|
|
def before_step(self, action):
|
|
assert not self.closed
|
|
|
|
if self.done:
|
|
raise error.ResetNeeded("Trying to step environment which is currently done. While the monitor is active, you cannot step beyond the end of an episode. Call 'env.reset()' to start the next episode.")
|
|
elif self.steps is None:
|
|
raise error.ResetNeeded("Trying to step an environment before reset. While the monitor is active, you must call 'env.reset()' before taking an initial step.")
|
|
|
|
def after_step(self, observation, reward, done, info):
|
|
self.steps += 1
|
|
self.rewards += reward
|
|
if done:
|
|
self.done = True
|
|
|
|
def before_reset(self):
|
|
assert not self.closed
|
|
|
|
self.done = False
|
|
if self.initial_reset_timestamp is None:
|
|
self.initial_reset_timestamp = time.time()
|
|
|
|
def after_reset(self, observation):
|
|
self.save_complete()
|
|
self.steps = 0
|
|
self.rewards = 0
|
|
|
|
def save_complete(self):
|
|
if self.steps is not None:
|
|
self.episode_lengths.append(self.steps)
|
|
self.episode_rewards.append(self.rewards)
|
|
self.timestamps.append(time.time())
|
|
|
|
def close(self):
|
|
self.save_complete()
|
|
self.flush()
|
|
self.closed = True
|
|
|
|
def flush(self):
|
|
if self.closed:
|
|
return
|
|
|
|
with atomic_write.atomic_write(self.path) as f:
|
|
json.dump({
|
|
'initial_reset_timestamp': self.initial_reset_timestamp,
|
|
'timestamps': self.timestamps,
|
|
'episode_lengths': self.episode_lengths,
|
|
'episode_rewards': self.episode_rewards,
|
|
}, f)
|