def init_engine(self, loader):
"""
Construct and save an FFCEngine from loader.
If loader is None, constructs a NoOpFFCEngine.
"""
if loader is not None:
self.engine = SimpleFFCEngine(
loader,
self.trading_environment.trading_days,
self.asset_finder,
)
else:
self.engine = NoOpFFCEngine()
def __init__(self, *args, **kwargs):
"""Initialize sids and other state variables.
:Arguments:
:Optional:
initialize : function
Function that is called with a single
argument at the begninning of the simulation.
handle_data : function
Function that is called with 2 arguments
(context and data) on every bar.
script : str
Algoscript that contains initialize and
handle_data function definition.
data_frequency : {'daily', 'minute'}
The duration of the bars.
capital_base : float <default: 1.0e5>
How much capital to start with.
instant_fill : bool <default: False>
Whether to fill orders immediately or on next bar.
asset_finder : An AssetFinder object
A new AssetFinder object to be used in this TradingEnvironment
asset_metadata: can be either:
- dict
- pandas.DataFrame
- object with 'read' property
If dict is provided, it must have the following structure:
* keys are the identifiers
* values are dicts containing the metadata, with the metadata
field name as the key
If pandas.DataFrame is provided, it must have the
following structure:
* column names must be the metadata fields
* index must be the different asset identifiers
* array contents should be the metadata value
If an object with a 'read' property is provided, 'read' must
return rows containing at least one of 'sid' or 'symbol' along
with the other metadata fields.
identifiers : List
Any asset identifiers that are not provided in the
asset_metadata, but will be traded by this TradingAlgorithm
"""
self.sources = []
# List of trading controls to be used to validate orders.
self.trading_controls = []
# List of account controls to be checked on each bar.
self.account_controls = []
self._recorded_vars = {}
self.namespace = kwargs.get('namespace', {})
self._platform = kwargs.pop('platform', 'zipline')
self.logger = None
self.benchmark_return_source = None
# default components for transact
self.slippage = VolumeShareSlippage()
self.commission = PerShare()
self.instant_fill = kwargs.pop('instant_fill', False)
# set the capital base
self.capital_base = kwargs.pop('capital_base', DEFAULT_CAPITAL_BASE)
self.sim_params = kwargs.pop('sim_params', None)
if self.sim_params is None:
self.sim_params = create_simulation_parameters(
capital_base=self.capital_base,
start=kwargs.pop('start', None),
end=kwargs.pop('end', None))
self.perf_tracker = PerformanceTracker(self.sim_params)
# Update the TradingEnvironment with the provided asset metadata
self.trading_environment = kwargs.pop('env',
TradingEnvironment.instance())
self.trading_environment.update_asset_finder(
asset_finder=kwargs.pop('asset_finder', None),
asset_metadata=kwargs.pop('asset_metadata', None),
identifiers=kwargs.pop('identifiers', None))
# Pull in the environment's new AssetFinder for quick reference
self.asset_finder = self.trading_environment.asset_finder
ffc_loader = kwargs.get('ffc_loader', None)
if ffc_loader is not None:
self.engine = SimpleFFCEngine(
ffc_loader,
self.trading_environment.trading_days,
self.asset_finder,
)
else:
self.engine = NoOpFFCEngine()
# Maps from name to Term
self._filters = {}
self._factors = {}
self._classifiers = {}
self.blotter = kwargs.pop('blotter', None)
if not self.blotter:
self.blotter = Blotter()
# Set the dt initally to the period start by forcing it to change
self.on_dt_changed(self.sim_params.period_start)
self.portfolio_needs_update = True
self.account_needs_update = True
self.performance_needs_update = True
self._portfolio = None
self._account = None
self.history_container_class = kwargs.pop(
'history_container_class',
HistoryContainer,
)
self.history_container = None
self.history_specs = {}
# If string is passed in, execute and get reference to
# functions.
self.algoscript = kwargs.pop('script', None)
self._initialize = None
self._before_trading_start = None
self._analyze = None
self.event_manager = EventManager()
if self.algoscript is not None:
filename = kwargs.pop('algo_filename', None)
if filename is None:
filename = '<string>'
code = compile(self.algoscript, filename, 'exec')
exec_(code, self.namespace)
self._initialize = self.namespace.get('initialize')
if 'handle_data' not in self.namespace:
raise ValueError('You must define a handle_data function.')
else:
self._handle_data = self.namespace['handle_data']
self._before_trading_start = \
self.namespace.get('before_trading_start')
# Optional analyze function, gets called after run
self._analyze = self.namespace.get('analyze')
elif kwargs.get('initialize') and kwargs.get('handle_data'):
if self.algoscript is not None:
raise ValueError('You can not set script and \
initialize/handle_data.')
self._initialize = kwargs.pop('initialize')
self._handle_data = kwargs.pop('handle_data')
self._before_trading_start = kwargs.pop('before_trading_start',
None)
self.event_manager.add_event(
zipline.utils.events.Event(
zipline.utils.events.Always(),
# We pass handle_data.__func__ to get the unbound method.
# We will explicitly pass the algorithm to bind it again.
self.handle_data.__func__,
),
prepend=True,
)
# If method not defined, NOOP
if self._initialize is None:
self._initialize = lambda x: None
# Alternative way of setting data_frequency for backwards
# compatibility.
if 'data_frequency' in kwargs:
self.data_frequency = kwargs.pop('data_frequency')
self._most_recent_data = None
# Prepare the algo for initialization
self.initialized = False
self.initialize_args = args
self.initialize_kwargs = kwargs