def __init__(self, *args, **kwargs):
"""
Environment kwargs applying logic::
if <engine> kwarg is given:
do not use default engine and strategy parameters;
ignore <strategy> kwarg and all strategy and engine-related kwargs.
else (no <engine>):
use default engine parameters;
if any engine-related kwarg is given:
override corresponding default parameter;
if <strategy> is given:
do not use default strategy parameters;
if any strategy related kwarg is given:
override corresponding strategy parameter;
else (no <strategy>):
use default strategy parameters;
if any strategy related kwarg is given:
override corresponding strategy parameter;
if <dataset> kwarg is given:
do not use default dataset parameters;
ignore dataset related kwargs;
else (no <dataset>):
use default dataset parameters;
if any dataset related kwarg is given:
override corresponding dataset parameter;
If any <other> kwarg is given:
override corresponding default parameter.
"""
#print("start backtrader")
# Parameters and default values:
self.params = dict(
# Backtrader engine mandatory parameters:
engine=dict(
start_cash=10.0, # initial trading capital.
broker_commission=
0.001, # trade execution commission, default is 0.1% of operation value.
fixed_stake=10, # single trade stake is fixed type by def.
),
# Dataset mandatory parameters:
dataset=dict(filename=None, ),
strategy=dict(state_shape=dict(), ),
render=dict(),
)
p2 = dict(
# Strategy related parameters:
# Observation state shape is dictionary of Gym spaces,
# at least should contain `raw_state` field.
# By convention first dimension of every Gym Box space is time embedding one;
# one can define any shape; should match env.observation_space.shape.
# observation space state min/max values,
# For `raw_state' - absolute min/max values from BTgymDataset will be used.
state_shape=dict(raw_state=spaces.Box(
shape=(10, 4),
low=-100,
high=100,
)),
drawdown_call=
None, # episode maximum drawdown threshold, default is 90% of initial value.
portfolio_actions=None,
# agent actions,
# should consist with BTgymStrategy order execution logic;
# defaults are: 0 - 'do nothing', 1 - 'buy', 2 - 'sell', 3 - 'close position'.
skip_frame=None,
# Number of environment steps to skip before returning next response,
# e.g. if set to 10 -- agent will interact with environment every 10th episode step;
# Every other step agent's action is assumed to be 'hold'.
# Note: INFO part of environment response is a list of all skipped frame's info's,
# i.e. [info[-9], info[-8], ..., info[0].
)
# Update self attributes, remove used kwargs:
for key in dir(self):
if key in kwargs.keys():
setattr(self, key, kwargs.pop(key))
self.metadata = {'render.modes': self.render_modes}
# Verbosity control:
if True: #self.log is None:
self.log = logging.getLogger('Env')
log_levels = [
(0, 'WARNING'),
(1, 'INFO'),
(2, 'DEBUG'),
]
for key, level in log_levels:
if key == self.verbose:
self.log.setLevel(level)
# Network parameters:
self.network_address += str(self.port)
self.data_network_address += str(self.data_port)
# Set server rendering:
if self.render_enabled:
self.renderer = BTgymRendering(self.metadata['render.modes'],
**kwargs)
else:
self.renderer = BTgymNullRendering()
self.log.info(
'Rendering disabled. Call to render() will return null-plug image.'
)
# Append logging:
self.renderer.log = self.log
# Update params -1: pull from renderer, remove used kwargs:
self.params['render'].update(self.renderer.params)
for key in self.params['render'].keys():
if key in kwargs.keys():
_ = kwargs.pop(key)
if self.data_master:
# DATASET preparation, only data_master executes this:
#
if self.dataset is not None:
# If BTgymDataset instance has been passed:
# do nothing.
msg = 'Custom Dataset class used.'
else:
# If no BTgymDataset has been passed,
# Make default dataset with given CSV file:
try:
os.path.isfile(str(self.params['dataset']['filename']))
except:
raise FileNotFoundError(
'Dataset source data file not specified/not found')
# Use kwargs to instantiate dataset:
self.dataset = BTgymDataset(**kwargs)
msg = 'Base Dataset class used.'
# Append logging:
self.dataset.log = self.log
# Update params -2: pull from dataset, remove used kwargs:
self.params['dataset'].update(self.dataset.params)
for key in self.params['dataset'].keys():
if key in kwargs.keys():
_ = kwargs.pop(key)
self.log.info(msg)
# Connect/Start data server (and get dataset statistic):
self.log.info('Connecting data_server...')
self._start_data_server()
self.log.info('...done.')
# ENGINE preparation:
# Update params -3: pull engine-related kwargs, remove used:
for key in self.params['engine'].keys():
if key in kwargs.keys():
self.params['engine'][key] = kwargs.pop(key)
if self.engine is not None:
# If full-blown bt.Cerebro() subclass has been passed:
# Update info:
msg = 'Custom Cerebro class used.'
self.strategy = msg
for key in self.params['engine'].keys():
self.params['engine'][key] = msg
# Note: either way, bt.observers.DrawDown observer [and logger] will be added to any BTgymStrategy instance
# by BTgymServer process at runtime.
else:
# Default configuration for Backtrader computational engine (Cerebro),
# if no bt.Cerebro() custom subclass has been passed,
# get base class Cerebro(), using kwargs on top of defaults:
self.engine = bt.Cerebro()
msg = 'Base Cerebro class used.'
# First, set STRATEGY configuration:
if self.strategy is not None:
# If custom strategy has been passed:
msg2 = 'Custom Strategy class used.'
else:
# Base class strategy :
self.strategy = BTgymBaseStrategy
msg2 = 'Base Strategy class used.'
# Add, using kwargs on top of defaults:
strat_idx = self.engine.addstrategy(self.strategy, **kwargs)
msg += ' ' + msg2
# Second, set Cerebro-level configuration:
self.engine.broker.setcash(self.params['engine']['start_cash'])
self.engine.broker.setcommission(
self.params['engine']['broker_commission'])
self.engine.addsizer(bt.sizers.SizerFix,
stake=self.params['engine']['fixed_stake'])
self.log.info(msg)
# Define observation space shape, minimum / maximum values and agent action space.
# Retrieve values from configured engine or...
# ...Update params -4:
# Pull strategy defaults to environment params dict :
for t_key, t_value in self.engine.strats[0][0][0].params._gettuple():
self.params['strategy'][t_key] = t_value
# Update it with values from strategy 'passed-to params':
for key, value in self.engine.strats[0][0][2].items():
self.params['strategy'][key] = value
# ... Push it all back (don't ask):
for key, value in self.params['strategy'].items():
self.engine.strats[0][0][2][key] = value
# For 'raw_state' min/max values,
# the only way is to infer from raw Dataset price values (we already got those from data_server):
if 'raw_state' in self.params['strategy']['state_shape'].keys():
# Exclude 'volume' from columns we count:
self.dataset_columns.remove('volume')
#print(self.params['strategy'])
#print('self.engine.strats[0][0][2]:', self.engine.strats[0][0][2])
#print('self.engine.strats[0][0][0].params:', self.engine.strats[0][0][0].params._gettuple())
# Override with absolute price min and max values:
self.params['strategy']['state_shape']['raw_state'].low =\
self.engine.strats[0][0][2]['state_shape']['raw_state'].low =\
np.zeros(self.params['strategy']['state_shape']['raw_state'].shape) +\
self.dataset_stat.loc['min', self.dataset_columns].min()
self.params['strategy']['state_shape']['raw_state'].high = \
self.engine.strats[0][0][2]['state_shape']['raw_state'].high = \
np.zeros(self.params['strategy']['state_shape']['raw_state'].shape) + \
self.dataset_stat.loc['max', self.dataset_columns].max()
self.log.info(
'Inferring `state_raw` high/low values form dataset: {:.6f} / {:.6f}.'
.format(
self.dataset_stat.loc['min', self.dataset_columns].min(),
self.dataset_stat.loc['max', self.dataset_columns].max()))
# Set observation space shape from engine/strategy parameters:
self.observation_space = DictSpace(
self.params['strategy']['state_shape'])
self.log.debug('Obs. shape: {}'.format(self.observation_space.spaces))
#self.log.debug('Obs. min:\n{}\nmax:\n{}'.format(self.observation_space.low, self.observation_space.high))
# Set action space and corresponding server messages:
self.action_space = spaces.Discrete(
len(self.params['strategy']['portfolio_actions']))
self.server_actions = self.params['strategy']['portfolio_actions']
# Finally:
self.server_response = None
self.env_response = None
# If instance is datamaster - it may or may not want to launch self BTgymServer (can do it later via reset);
# else it always need to launch it:
#if not self.data_master:
self._start_server()
self.closed = False
self.log.info('Environment is ready.')
def __init__(self, **kwargs):
"""
Keyword Args:
filename=None (str, list): csv data file.
**datafeed_args (any): any datafeed-related args, passed through to
default btgym.datafeed class.
dataset=None (btgym.datafeed): BTgymDataDomain instance,
overrides `filename` or any other datafeed-related args.
strategy=None (btgym.startegy): strategy to be used by `engine`, any subclass of
btgym.strategy.base.BTgymBaseStrateg
engine=None (bt.Cerebro): environment simulation engine, any bt.Cerebro subclass,
overrides `strategy` arg.
network_address=`tcp://127.0.0.1:` (str): BTGym_server address.
port=5500 (int): network port to use for server - API_shell communication.
data_master=True (bool): let this environment control over data_server;
data_network_address=`tcp://127.0.0.1:` (str): data_server address.
data_port=4999 (int): network port to use for server -- data_server communication.
connect_timeout=60 (int): server connection timeout in seconds.
render_enabled=True (bool): enable rendering for this environment;
render_modes=['human', 'episode'] (list): `episode` - plotted episode results;
`human` - raw_state observation.
**render_args (any): any render-related args, passed through to renderer class.
verbose=0 (int): verbosity mode, {0 - WARNING, 1 - INFO, 2 - DEBUG}
log_level=None (int): logbook level {DEBUG=10, INFO=11, NOTICE=12, WARNING=13},
overrides `verbose` arg;
log=None (logbook.Logger): external logbook logger,
overrides `log_level` and `verbose` args.
task=0 (int): environment id
Environment kwargs applying logic::
if <engine> kwarg is given:
do not use default engine and strategy parameters;
ignore <strategy> kwarg and all strategy and engine-related kwargs.
else (no <engine>):
use default engine parameters;
if any engine-related kwarg is given:
override corresponding default parameter;
if <strategy> is given:
do not use default strategy parameters;
if any strategy related kwarg is given:
override corresponding strategy parameter;
else (no <strategy>):
use default strategy parameters;
if any strategy related kwarg is given:
override corresponding strategy parameter;
if <dataset> kwarg is given:
do not use default dataset parameters;
ignore dataset related kwargs;
else (no <dataset>):
use default dataset parameters;
if any dataset related kwarg is given:
override corresponding dataset parameter;
If any <other> kwarg is given:
override corresponding default parameter.
"""
# Parameters and default values:
self.params = dict(
# Backtrader engine mandatory parameters:
engine=dict(
start_cash=10.0, # initial trading capital.
broker_commission=
0.001, # trade execution commission, default is 0.1% of operation value.
fixed_stake=10, # single trade stake is fixed type by def.
),
# Dataset mandatory parameters:
dataset=dict(filename=None, ),
strategy=dict(state_shape=dict(), ),
render=dict(),
)
p2 = dict( # IS HERE FOR REFERENCE ONLY
# Strategy related parameters:
# Observation state shape is dictionary of Gym spaces,
# at least should contain `raw_state` field.
# By convention first dimension of every Gym Box space is time embedding one;
# one can define any shape; should match env.observation_space.shape.
# observation space state min/max values,
# For `raw_state' - absolute min/max values from BTgymDataset will be used.
state_shape=dict(raw_state=spaces.Box(
shape=(10, 4), low=-100, high=100, dtype=np.float32)),
drawdown_call=
None, # episode maximum drawdown threshold, default is 90% of initial value.
portfolio_actions=None,
# agent actions,
# should consist with BTgymStrategy order execution logic;
# defaults are: 0 - 'do nothing', 1 - 'buy', 2 - 'sell', 3 - 'close position'.
skip_frame=None,
# Number of environment steps to skip before returning next response,
# e.g. if set to 10 -- agent will interact with environment every 10th episode step;
# Every other step agent's action is assumed to be 'hold'.
# Note: INFO part of environment response is a list of all skipped frame's info's,
# i.e. [info[-9], info[-8], ..., info[0].
)
# Update self attributes, remove used kwargs:
for key in dir(self):
if key in kwargs.keys():
setattr(self, key, kwargs.pop(key))
self.metadata = {'render.modes': self.render_modes}
# Logging and verbosity control:
if self.log is None:
StreamHandler(sys.stdout).push_application()
if self.log_level is None:
log_levels = [(0, NOTICE), (1, INFO), (2, DEBUG)]
self.log_level = WARNING
for key, value in log_levels:
if key == self.verbose:
self.log_level = value
self.log = Logger('BTgymAPIshell_{}'.format(self.task),
level=self.log_level)
# Network parameters:
self.network_address += str(self.port)
self.data_network_address += str(self.data_port)
# Set server rendering:
if self.render_enabled:
self.renderer = BTgymRendering(self.metadata['render.modes'],
log_level=self.log_level,
**kwargs)
else:
self.renderer = BTgymNullRendering()
self.log.info(
'Rendering disabled. Call to render() will return null-plug image.'
)
# Append logging:
self.renderer.log = self.log
# Update params -1: pull from renderer, remove used kwargs:
self.params['render'].update(self.renderer.params)
for key in self.params['render'].keys():
if key in kwargs.keys():
_ = kwargs.pop(key)
if self.data_master:
# DATASET preparation, only data_master executes this:
#
if self.dataset is not None:
# If BTgymDataset instance has been passed:
# do nothing.
msg = 'Custom Dataset class used.'
else:
# If no BTgymDataset has been passed,
# Make default dataset with given CSV file:
try:
os.path.isfile(str(self.params['dataset']['filename']))
except:
raise FileNotFoundError(
'Dataset source data file not specified/not found')
# Use kwargs to instantiate dataset:
self.dataset = BTgymDataset(**kwargs)
msg = 'Base Dataset class used.'
# Append logging:
self.dataset.set_logger(self.log_level, self.task)
# Update params -2: pull from dataset, remove used kwargs:
self.params['dataset'].update(self.dataset.params)
for key in self.params['dataset'].keys():
if key in kwargs.keys():
_ = kwargs.pop(key)
self.log.info(msg)
# Connect/Start data server (and get dataset statistic):
self.log.info('Connecting data_server...')
self._start_data_server()
self.log.info('...done.')
# ENGINE preparation:
# Update params -3: pull engine-related kwargs, remove used:
for key in self.params['engine'].keys():
if key in kwargs.keys():
self.params['engine'][key] = kwargs.pop(key)
if self.engine is not None:
# If full-blown bt.Cerebro() subclass has been passed:
# Update info:
msg = 'Custom Cerebro class used.'
self.strategy = msg
for key in self.params['engine'].keys():
self.params['engine'][key] = msg
# Note: either way, bt.observers.DrawDown observer [and logger] will be added to any BTgymStrategy instance
# by BTgymServer process at runtime.
else:
# Default configuration for Backtrader computational engine (Cerebro),
# if no bt.Cerebro() custom subclass has been passed,
# get base class Cerebro(), using kwargs on top of defaults:
self.engine = bt.Cerebro()
msg = 'Base Cerebro class used.'
# First, set STRATEGY configuration:
if self.strategy is not None:
# If custom strategy has been passed:
msg2 = 'Custom Strategy class used.'
else:
# Base class strategy :
self.strategy = BTgymBaseStrategy
msg2 = 'Base Strategy class used.'
# Add, using kwargs on top of defaults:
#self.log.debug('kwargs for strategy: {}'.format(kwargs))
strat_idx = self.engine.addstrategy(self.strategy, **kwargs)
msg += ' ' + msg2
# Second, set Cerebro-level configuration:
self.engine.broker.setcash(self.params['engine']['start_cash'])
self.engine.broker.setcommission(
self.params['engine']['broker_commission'])
self.engine.addsizer(bt.sizers.SizerFix,
stake=self.params['engine']['fixed_stake'])
self.log.info(msg)
# Define observation space shape, minimum / maximum values and agent action space.
# Retrieve values from configured engine or...
# ...Update params -4:
# Pull strategy defaults to environment params dict :
for t_key, t_value in self.engine.strats[0][0][0].params._gettuple():
self.params['strategy'][t_key] = t_value
# Update it with values from strategy 'passed-to params':
for key, value in self.engine.strats[0][0][2].items():
self.params['strategy'][key] = value
# ... Push it all back (don't ask):
for key, value in self.params['strategy'].items():
self.engine.strats[0][0][2][key] = value
# For 'raw_state' min/max values,
# the only way is to infer from raw Dataset price values (we already got those from data_server):
if 'raw_state' in self.params['strategy']['state_shape'].keys():
# Exclude 'volume' from columns we count:
self.dataset_columns.remove('volume')
#print(self.params['strategy'])
#print('self.engine.strats[0][0][2]:', self.engine.strats[0][0][2])
#print('self.engine.strats[0][0][0].params:', self.engine.strats[0][0][0].params._gettuple())
# Override with absolute price min and max values:
self.params['strategy']['state_shape']['raw_state'].low =\
self.engine.strats[0][0][2]['state_shape']['raw_state'].low =\
np.zeros(self.params['strategy']['state_shape']['raw_state'].shape) +\
self.dataset_stat.loc['min', self.dataset_columns].min()
self.params['strategy']['state_shape']['raw_state'].high = \
self.engine.strats[0][0][2]['state_shape']['raw_state'].high = \
np.zeros(self.params['strategy']['state_shape']['raw_state'].shape) + \
self.dataset_stat.loc['max', self.dataset_columns].max()
self.log.info(
'Inferring `state_raw` high/low values form dataset: {:.6f} / {:.6f}.'
.format(
self.dataset_stat.loc['min', self.dataset_columns].min(),
self.dataset_stat.loc['max', self.dataset_columns].max()))
# Set observation space shape from engine/strategy parameters:
self.observation_space = DictSpace(
self.params['strategy']['state_shape'])
self.log.debug('Obs. shape: {}'.format(self.observation_space.spaces))
#self.log.debug('Obs. min:\n{}\nmax:\n{}'.format(self.observation_space.low, self.observation_space.high))
# Set action space and corresponding server messages:
self.action_space = spaces.Discrete(
len(self.params['strategy']['portfolio_actions']))
self.server_actions = self.params['strategy']['portfolio_actions']
# Finally:
self.server_response = None
self.env_response = None
#if not self.data_master:
self._start_server()
self.closed = False
self.log.info('Environment is ready.')
def __init__(self, engine, dataset=None, **kwargs):
"""
This class requires dataset, strategy, engine instances to be passed explicitly.
Args:
dataset(btgym.datafeed): BTgymDataDomain instance;
engine(bt.Cerebro): environment simulation engine, any bt.Cerebro subclass,
Keyword Args:
network_address=`tcp://127.0.0.1:` (str): BTGym_server address.
port=5500 (int): network port to use for server - API_shell communication.
data_master=True (bool): let this environment control over data_server;
data_network_address=`tcp://127.0.0.1:` (str): data_server address.
data_port=4999 (int): network port to use for server -- data_server communication.
connect_timeout=20 (int): server connection timeout in seconds.
render_enabled=True (bool): enable rendering for this environment;
render_modes=['human', 'episode'] (list): `episode` - plotted episode results;
`human` - raw_state observation.
**render_args (any): any render-related args, passed through to renderer class.
verbose=0 (int): verbosity mode, {0 - WARNING, 1 - INFO, 2 - DEBUG}
log_level=None (int): logbook level {DEBUG=10, INFO=11, NOTICE=12, WARNING=13},
overrides `verbose` arg;
log=None (logbook.Logger): external logbook logger,
overrides `log_level` and `verbose` args.
task=0 (int): environment id
"""
self.dataset = dataset
self.engine = engine
# Parameters and default values:
self.params = dict(
engine={},
dataset={},
strategy={},
render={},
)
# Update self attributes, remove used kwargs:
for key in dir(self):
if key in kwargs.keys():
setattr(self, key, kwargs.pop(key))
self.metadata = {'render.modes': self.render_modes}
# Logging and verbosity control:
if self.log is None:
StreamHandler(sys.stdout).push_application()
if self.log_level is None:
log_levels = [(0, NOTICE), (1, INFO), (2, DEBUG)]
self.log_level = WARNING
for key, value in log_levels:
if key == self.verbose:
self.log_level = value
self.log = Logger('BTgymMultiDataShell_{}'.format(self.task), level=self.log_level)
# Network parameters:
self.network_address += str(self.port)
self.data_network_address += str(self.data_port)
# Set server rendering:
if self.render_enabled:
self.renderer = BTgymRendering(self.metadata['render.modes'], log_level=self.log_level, **kwargs)
else:
self.renderer = BTgymNullRendering()
self.log.info('Rendering disabled. Call to render() will return null-plug image.')
# Append logging:
self.renderer.log = self.log
# Update params -1: pull from renderer, remove used kwargs:
self.params['render'].update(self.renderer.params)
for key in self.params['render'].keys():
if key in kwargs.keys():
_ = kwargs.pop(key)
if self.data_master:
try:
assert self.dataset is not None
except AssertionError:
msg = 'Dataset instance shoud be provided for data_master environment.'
self.log.error(msg)
raise ValueError(msg)
# Append logging:
self.dataset.set_logger(self.log_level, self.task)
# Update params -2: pull from dataset, remove used kwargs:
self.params['dataset'].update(self.dataset.params)
for key in self.params['dataset'].keys():
if key in kwargs.keys():
_ = kwargs.pop(key)
# Connect/Start data server (and get dataset statistic):
self.log.info('Connecting data_server...')
self._start_data_server()
self.log.info('...done.')
# After starting data-server we have self.assets attribute, dataset statisitc etc. filled.
# Define observation space shape, minimum / maximum values and agent action space.
# Retrieve values from configured engine or...
# ...Update params -4:
# Pull strategy defaults to environment params dict :
for t_key, t_value in self.engine.strats[0][0][0].params._gettuple():
self.params['strategy'][t_key] = t_value
# Update it with values from strategy 'passed-to params':
for key, value in self.engine.strats[0][0][2].items():
self.params['strategy'][key] = value
self.asset_names = self.params['strategy']['asset_names']
self.server_actions = {name: self.params['strategy']['portfolio_actions'] for name in self.asset_names}
self.cash_name = self.params['strategy']['cash_name']
self.params['strategy']['initial_action'] = self.get_initial_action()
self.params['strategy']['initial_portfolio_action'] = self.get_initial_portfolio_action()
try:
assert set(self.asset_names).issubset(set(self.data_lines_names))
except AssertionError:
msg = 'Assets names should be subset of data_lines names, but got: assets: {}, data_lines: {}'.format(
set(self.asset_names), set(self.data_lines_names)
)
self.log.error(msg)
raise ValueError(msg)
# ... Push it all back (don't ask):
for key, value in self.params['strategy'].items():
self.engine.strats[0][0][2][key] = value
# For 'raw_state' min/max values,
# the only way is to infer from raw Dataset price values (we already got those from data_server):
if 'raw_state' in self.params['strategy']['state_shape'].keys():
# Exclude 'volume' from columns we count:
self.dataset_columns.remove('volume')
# print(self.params['strategy'])
# print('self.engine.strats[0][0][2]:', self.engine.strats[0][0][2])
# print('self.engine.strats[0][0][0].params:', self.engine.strats[0][0][0].params._gettuple())
# Override with absolute price min and max values:
self.params['strategy']['state_shape']['raw_state'].low = \
self.engine.strats[0][0][2]['state_shape']['raw_state'].low = \
np.zeros(self.params['strategy']['state_shape']['raw_state'].shape) + \
self.dataset_stat.loc['min', self.dataset_columns].min()
self.params['strategy']['state_shape']['raw_state'].high = \
self.engine.strats[0][0][2]['state_shape']['raw_state'].high = \
np.zeros(self.params['strategy']['state_shape']['raw_state'].shape) + \
self.dataset_stat.loc['max', self.dataset_columns].max()
self.log.info('Inferring `state_raw` high/low values form dataset: {:.6f} / {:.6f}.'.
format(self.dataset_stat.loc['min', self.dataset_columns].min(),
self.dataset_stat.loc['max', self.dataset_columns].max()))
# Set observation space shape from engine/strategy parameters:
self.observation_space = DictSpace(self.params['strategy']['state_shape'])
self.log.debug('Obs. shape: {}'.format(self.observation_space.spaces))
# Set action space and corresponding server messages:
self.action_space = ActionDictSpace(
base_actions=self.params['strategy']['portfolio_actions'],
assets=self.asset_names
)
self.log.debug('Act. space shape: {}'.format(self.action_space.spaces))
# Finally:
self.server_response = None
self.env_response = None
# if not self.data_master:
self._start_server()
self.closed = False
self.log.info('Environment is ready.')