API#

Running a Backtest#

The function run_algorithm() creates an instance of TradingAlgorithm that represents a trading strategy and parameters to execute the strategy.

zipline.run_algorithm(...)[source]#

Run a trading algorithm.

Parameters:

start (datetime) – The start date of the backtest.
end (datetime) – The end date of the backtest..
initialize (callable[context -> None]) – The initialize function to use for the algorithm. This is called once at the very begining of the backtest and should be used to set up any state needed by the algorithm.
capital_base (float) – The starting capital for the backtest.
handle_data (callable[(context, BarData) -> None], optional) – The handle_data function to use for the algorithm. This is called every minute when data_frequency == 'minute' or every day when data_frequency == 'daily'.
before_trading_start (callable[(context, BarData) -> None], optional) – The before_trading_start function for the algorithm. This is called once before each trading day (after initialize on the first day).
analyze (callable[(context, pd.DataFrame) -> None], optional) – The analyze function to use for the algorithm. This function is called once at the end of the backtest and is passed the context and the performance data.
data_frequency ({'daily', 'minute'}, optional) – The data frequency to run the algorithm at.
bundle (str, optional) – The name of the data bundle to use to load the data to run the backtest with. This defaults to ‘quantopian-quandl’.
bundle_timestamp (datetime, optional) – The datetime to lookup the bundle data for. This defaults to the current time.
trading_calendar (TradingCalendar, optional) – The trading calendar to use for your backtest.
metrics_set (iterable[Metric] or str, optional) – The set of metrics to compute in the simulation. If a string is passed, resolve the set with zipline.finance.metrics.load().
benchmark_returns (pd.Series, optional) – Series of returns to use as the benchmark.
default_extension (bool, optional) – Should the default zipline extension be loaded. This is found at $ZIPLINE_ROOT/extension.py
extensions (iterable[str], optional) – The names of any other extensions to load. Each element may either be a dotted module path like a.b.c or a path to a python file ending in .py like a/b/c.py.
strict_extensions (bool, optional) – Should the run fail if any extensions fail to load. If this is false, a warning will be raised instead.
environ (mapping[str -> str], optional) – The os environment to use. Many extensions use this to get parameters. This defaults to os.environ.
blotter (str or zipline.finance.blotter.Blotter, optional) – Blotter to use with this algorithm. If passed as a string, we look for a blotter construction function registered with zipline.extensions.register and call it with no parameters. Default is a zipline.finance.blotter.SimulationBlotter that never cancels orders.

Returns:

perf – The daily performance of the algorithm.

Return type:

pd.DataFrame

See also

zipline.data.bundles.bundles: The available data bundles.

Trading Algorithm API#

The following methods are available for use in the initialize, handle_data, and before_trading_start API functions.

In all listed functions, the self argument refers to the currently executing TradingAlgorithm instance.

Data Object#

class zipline.protocol.BarData#

Provides methods for accessing minutely and daily price/volume data from Algorithm API functions.

Also provides utility methods to determine if an asset is alive, and if it has recent trade data.

An instance of this object is passed as data to handle_data() and before_trading_start().

Parameters:

data_portal (DataPortal) – Provider for bar pricing data.
simulation_dt_func (callable) – Function which returns the current simulation time. This is usually bound to a method of TradingSimulation.
data_frequency ({'minute', 'daily'}) – The frequency of the bar data; i.e. whether the data is daily or minute bars
restrictions (zipline.finance.asset_restrictions.Restrictions) – Object that combines and returns restricted list information from multiple sources

can_trade()#

For the given asset or iterable of assets, returns True if all of the following are true:

The asset is alive for the session of the current simulation time (if current simulation time is not a market minute, we use the next session).
The asset’s exchange is open at the current simulation time or at the simulation calendar’s next market minute.
There is a known last price for the asset.

Parameters:: assets (zipline.assets.Asset or iterable of zipline.assets.Asset) – Asset(s) for which tradability should be determined.

Notes

The second condition above warrants some further explanation:

If the asset’s exchange calendar is identical to the simulation calendar, then this condition always returns True.
If there are market minutes in the simulation calendar outside of this asset’s exchange’s trading hours (for example, if the simulation is running on the CMES calendar but the asset is MSFT, which trades on the NYSE), during those minutes, this condition will return False (for example, 3:15 am Eastern on a weekday, during which the CMES is open but the NYSE is closed).

Returns:: can_trade – Bool or series of bools indicating whether the requested asset(s) can be traded in the current minute.
Return type:: bool or pd.Series[bool]

current()#

Returns the “current” value of the given fields for the given assets at the current simulation time.

Parameters:

assets (zipline.assets.Asset or iterable of zipline.assets.Asset) – The asset(s) for which data is requested.
fields (str or iterable[str].) – Requested data field(s). Valid field names are: “price”, “last_traded”, “open”, “high”, “low”, “close”, and “volume”.

Returns:

current_value – See notes below.

Return type:

Scalar, pandas Series, or pandas DataFrame.

Notes

The return type of this function depends on the types of its inputs:

If a single asset and a single field are requested, the returned value is a scalar (either a float or a pd.Timestamp depending on the field).
If a single asset and a list of fields are requested, the returned value is a pd.Series whose indices are the requested fields.
If a list of assets and a single field are requested, the returned value is a pd.Series whose indices are the assets.
If a list of assets and a list of fields are requested, the returned value is a pd.DataFrame. The columns of the returned frame will be the requested fields, and the index of the frame will be the requested assets.

The values produced for fields are as follows:

Requesting “price” produces the last known close price for the asset, forward-filled from an earlier minute if there is no trade this minute. If there is no last known value (either because the asset has never traded, or because it has delisted) NaN is returned. If a value is found, and we had to cross an adjustment boundary (split, dividend, etc) to get it, the value is adjusted to the current simulation time before being returned.
Requesting “open”, “high”, “low”, or “close” produces the open, high, low, or close for the current minute. If no trades occurred this minute, NaN is returned.
Requesting “volume” produces the trade volume for the current minute. If no trades occurred this minute, 0 is returned.
Requesting “last_traded” produces the datetime of the last minute in which the asset traded, even if the asset has stopped trading. If there is no last known value, pd.NaT is returned.

If the current simulation time is not a valid market time for an asset, we use the most recent market close instead.

history()#

Returns a trailing window of length bar_count with data for the given assets, fields, and frequency, adjusted for splits, dividends, and mergers as of the current simulation time.

The semantics for missing data are identical to the ones described in the notes for current().

Parameters:

assets (zipline.assets.Asset or iterable of zipline.assets.Asset) – The asset(s) for which data is requested.
fields (string or iterable of string.) – Requested data field(s). Valid field names are: “price”, “last_traded”, “open”, “high”, “low”, “close”, and “volume”.
bar_count (int) – Number of data observations requested.
frequency (str) – String indicating whether to load daily or minutely data observations. Pass ‘1m’ for minutely data, ‘1d’ for daily data.

Returns:

history – See notes below.

Return type:

pd.Series or pd.DataFrame or pd.Panel

Notes

The return type of this function depends on the types of assets and fields:

If a single asset and a single field are requested, the returned value is a pd.Series of length bar_count whose index is pd.DatetimeIndex.
If a single asset and multiple fields are requested, the returned value is a pd.DataFrame with shape (bar_count, len(fields)). The frame’s index will be a pd.DatetimeIndex, and its columns will be fields.
If multiple assets and a single field are requested, the returned value is a pd.DataFrame with shape (bar_count, len(assets)). The frame’s index will be a pd.DatetimeIndex, and its columns will be assets.
If multiple assets and multiple fields are requested, the returned value is a pd.DataFrame with a pd.MultiIndex containing pairs of pd.DatetimeIndex, and assets, while the columns while contain the field(s). It has shape (bar_count * len(assets), len(fields)). The names of the pd.MultiIndex are
- date if frequency == ‘1d’`` or date_time if frequency == ‘1m``, and
- asset

If the current simulation time is not a valid market time, we use the last market close instead.

is_stale()#

For the given asset or iterable of assets, returns True if the asset is alive and there is no trade data for the current simulation time.

If the asset has never traded, returns False.

If the current simulation time is not a valid market time, we use the current time to check if the asset is alive, but we use the last market minute/day for the trade data check.

Parameters:: assets (zipline.assets.Asset or iterable of zipline.assets.Asset) – Asset(s) for which staleness should be determined.
Returns:: is_stale – Bool or series of bools indicating whether the requested asset(s) are stale.
Return type:: bool or pd.Series[bool]

Scheduling Functions#

zipline.api.schedule_function(self, func, date_rule=None, time_rule=None, half_days=True, calendar=None)#

Schedule a function to be called repeatedly in the future.

Parameters:

func (callable) – The function to execute when the rule is triggered. func should have the same signature as handle_data.
date_rule (zipline.utils.events.EventRule, optional) – Rule for the dates on which to execute func. If not passed, the function will run every trading day.
time_rule (zipline.utils.events.EventRule, optional) – Rule for the time at which to execute func. If not passed, the function will execute at the end of the first market minute of the day.
half_days (bool, optional) – Should this rule fire on half days? Default is True.
calendar (Sentinel, optional) – Calendar used to compute rules that depend on the trading calendar.

class zipline.api.date_rules[source]#

Factories for date-based schedule_function() rules.

See also

schedule_function()

static every_day()[source]#

Create a rule that triggers every day.

Returns:: rule
Return type:: zipline.utils.events.EventRule

static month_end(days_offset=0)[source]#

Create a rule that triggers a fixed number of trading days before the end of each month.

Parameters:: days_offset (int, optional) – Number of trading days prior to month end to trigger. Default is 0, i.e., trigger on the last day of the month.
Returns:: rule
Return type:: zipline.utils.events.EventRule

static month_start(days_offset=0)[source]#

Create a rule that triggers a fixed number of trading days after the start of each month.

Parameters:: days_offset (int, optional) – Number of trading days to wait before triggering each month. Default is 0, i.e., trigger on the first trading day of the month.
Returns:: rule
Return type:: zipline.utils.events.EventRule

static week_end(days_offset=0)[source]#

Create a rule that triggers a fixed number of trading days before the end of each week.

Parameters:: days_offset (int, optional) – Number of trading days prior to week end to trigger. Default is 0, i.e., trigger on the last trading day of the week.

static week_start(days_offset=0)[source]#

Create a rule that triggers a fixed number of trading days after the start of each week.

Parameters:: days_offset (int, optional) – Number of trading days to wait before triggering each week. Default is 0, i.e., trigger on the first trading day of the week.

class zipline.api.time_rules[source]#

Factories for time-based schedule_function() rules.

See also

schedule_function()

every_minute#: alias of Always

static market_close(offset=None, hours=None, minutes=None)[source]#

Create a rule that triggers at a fixed offset from market close.

The offset can be specified either as a datetime.timedelta, or as a number of hours and minutes.

Parameters:

offset (datetime.timedelta, optional) – If passed, the offset from market close at which to trigger. Must be at least 1 minute.
hours (int, optional) – If passed, number of hours to wait before market close.
minutes (int, optional) – If passed, number of minutes to wait before market close.

Returns:

rule

Return type:

zipline.utils.events.EventRule

Notes

If no arguments are passed, the default offset is one minute before market close.

If offset is passed, hours and minutes must not be passed. Conversely, if either hours or minutes are passed, offset must not be passed.

static market_open(offset=None, hours=None, minutes=None)[source]#

Create a rule that triggers at a fixed offset from market open.

The offset can be specified either as a datetime.timedelta, or as a number of hours and minutes.

Parameters:

offset (datetime.timedelta, optional) – If passed, the offset from market open at which to trigger. Must be at least 1 minute.
hours (int, optional) – If passed, number of hours to wait after market open.
minutes (int, optional) – If passed, number of minutes to wait after market open.

Returns:

rule

Return type:

zipline.utils.events.EventRule

Notes

If no arguments are passed, the default offset is one minute after market open.

If offset is passed, hours and minutes must not be passed. Conversely, if either hours or minutes are passed, offset must not be passed.

Orders#

zipline.api.order(self, asset, amount, limit_price=None, stop_price=None, style=None)#

Place an order for a fixed number of shares.

Parameters:

asset (Asset) – The asset to be ordered.
amount (int) – The amount of shares to order. If amount is positive, this is the number of shares to buy or cover. If amount is negative, this is the number of shares to sell or short.
limit_price (float, optional) – The limit price for the order.
stop_price (float, optional) – The stop price for the order.
style (ExecutionStyle, optional) – The execution style for the order.

Returns:

order_id – The unique identifier for this order, or None if no order was placed.

Return type:

str or None

Notes

The limit_price and stop_price arguments provide shorthands for passing common execution styles. Passing limit_price=N is equivalent to style=LimitOrder(N). Similarly, passing stop_price=M is equivalent to style=StopOrder(M), and passing limit_price=N and stop_price=M is equivalent to style=StopLimitOrder(N, M). It is an error to pass both a style and limit_price or stop_price.

zipline.api.order_value(self, asset, value, limit_price=None, stop_price=None, style=None)#

Place an order for a fixed amount of money.

Equivalent to order(asset, value / data.current(asset, 'price')).

Parameters:

asset (Asset) – The asset to be ordered.
value (float) – Amount of value of asset to be transacted. The number of shares bought or sold will be equal to value / current_price.
limit_price (float, optional) – Limit price for the order.
stop_price (float, optional) – Stop price for the order.
style (ExecutionStyle) – The execution style for the order.

Returns:

order_id – The unique identifier for this order.

Return type:

str

Notes

See zipline.api.order() for more information about limit_price, stop_price, and style

zipline.api.order_percent(self, asset, percent, limit_price=None, stop_price=None, style=None)#

Place an order in the specified asset corresponding to the given percent of the current portfolio value.

Parameters:

asset (Asset) – The asset that this order is for.
percent (float) – The percentage of the portfolio value to allocate to asset. This is specified as a decimal, for example: 0.50 means 50%.
limit_price (float, optional) – The limit price for the order.
stop_price (float, optional) – The stop price for the order.
style (ExecutionStyle) – The execution style for the order.

Returns:

order_id – The unique identifier for this order.

Return type:

str

Notes

See zipline.api.order() for more information about limit_price, stop_price, and style

zipline.api.order_target(self, asset, target, limit_price=None, stop_price=None, style=None)#

Place an order to adjust a position to a target number of shares. If the position doesn’t already exist, this is equivalent to placing a new order. If the position does exist, this is equivalent to placing an order for the difference between the target number of shares and the current number of shares.

Parameters:

asset (Asset) – The asset that this order is for.
target (int) – The desired number of shares of asset.
limit_price (float, optional) – The limit price for the order.
stop_price (float, optional) – The stop price for the order.
style (ExecutionStyle) – The execution style for the order.

Returns:

order_id – The unique identifier for this order.

Return type:

str

Notes

order_target does not take into account any open orders. For example:

order_target(sid(0), 10)
order_target(sid(0), 10)

This code will result in 20 shares of sid(0) because the first call to order_target will not have been filled when the second order_target call is made.

See zipline.api.order() for more information about limit_price, stop_price, and style

zipline.api.order_target_value(self, asset, target, limit_price=None, stop_price=None, style=None)#

Place an order to adjust a position to a target value. If the position doesn’t already exist, this is equivalent to placing a new order. If the position does exist, this is equivalent to placing an order for the difference between the target value and the current value. If the Asset being ordered is a Future, the ‘target value’ calculated is actually the target exposure, as Futures have no ‘value’.

Parameters:

asset (Asset) – The asset that this order is for.
target (float) – The desired total value of asset.
limit_price (float, optional) – The limit price for the order.
stop_price (float, optional) – The stop price for the order.
style (ExecutionStyle) – The execution style for the order.

Returns:

order_id – The unique identifier for this order.

Return type:

str

Notes

order_target_value does not take into account any open orders. For example:

order_target_value(sid(0), 10)
order_target_value(sid(0), 10)

This code will result in 20 dollars of sid(0) because the first call to order_target_value will not have been filled when the second order_target_value call is made.

See zipline.api.order() for more information about limit_price, stop_price, and style

zipline.api.order_target_percent(self, asset, target, limit_price=None, stop_price=None, style=None)#

Place an order to adjust a position to a target percent of the current portfolio value. If the position doesn’t already exist, this is equivalent to placing a new order. If the position does exist, this is equivalent to placing an order for the difference between the target percent and the current percent.

Parameters:

asset (Asset) – The asset that this order is for.
target (float) – The desired percentage of the portfolio value to allocate to asset. This is specified as a decimal, for example: 0.50 means 50%.
limit_price (float, optional) – The limit price for the order.
stop_price (float, optional) – The stop price for the order.
style (ExecutionStyle) – The execution style for the order.

Returns:

order_id – The unique identifier for this order.

Return type:

str

Notes

order_target_value does not take into account any open orders. For example:

order_target_percent(sid(0), 10)
order_target_percent(sid(0), 10)

This code will result in 20% of the portfolio being allocated to sid(0) because the first call to order_target_percent will not have been filled when the second order_target_percent call is made.

See zipline.api.order() for more information about limit_price, stop_price, and style

class zipline.finance.execution.ExecutionStyle[source]#

Base class for order execution styles.

property exchange#: The exchange to which this order should be routed.

abstract get_limit_price(is_buy)[source]#: Get the limit price for this order. Returns either None or a numerical value >= 0.

abstract get_stop_price(is_buy)[source]#: Get the stop price for this order. Returns either None or a numerical value >= 0.

class zipline.finance.execution.MarketOrder(exchange=None)[source]#

Execution style for orders to be filled at current market price.

This is the default for orders placed with order().

class zipline.finance.execution.LimitOrder(limit_price, asset=None, exchange=None)[source]#

Execution style for orders to be filled at a price equal to or better than a specified limit price.

Parameters:: limit_price (float) – Maximum price for buys, or minimum price for sells, at which the order should be filled.

class zipline.finance.execution.StopOrder(stop_price, asset=None, exchange=None)[source]#

Execution style representing a market order to be placed if market price reaches a threshold.

Parameters:: stop_price (float) – Price threshold at which the order should be placed. For sells, the order will be placed if market price falls below this value. For buys, the order will be placed if market price rises above this value.

class zipline.finance.execution.StopLimitOrder(limit_price, stop_price, asset=None, exchange=None)[source]#

Execution style representing a limit order to be placed if market price reaches a threshold.

Parameters:

limit_price (float) – Maximum price for buys, or minimum price for sells, at which the order should be filled, if placed.
stop_price (float) – Price threshold at which the order should be placed. For sells, the order will be placed if market price falls below this value. For buys, the order will be placed if market price rises above this value.

zipline.api.get_order(self, order_id)#

Lookup an order based on the order id returned from one of the order functions.

Parameters:: order_id (str) – The unique identifier for the order.
Returns:: order – The order object.
Return type:: Order

zipline.api.get_open_orders(self, asset=None)#

Retrieve all of the current open orders.

Parameters:: asset (Asset) – If passed and not None, return only the open orders for the given asset instead of all open orders.
Returns:: open_orders – If no asset is passed this will return a dict mapping Assets to a list containing all the open orders for the asset. If an asset is passed then this will return a list of the open orders for this asset.
Return type:: dict[list[Order]] or list[Order]

zipline.api.cancel_order(self, order_param)#

Cancel an open order.

Parameters:: order_param (str or Order) – The order_id or order object to cancel.

Order Cancellation Policies#

zipline.api.set_cancel_policy(self, cancel_policy)#

Sets the order cancellation policy for the simulation.

Parameters:: cancel_policy (CancelPolicy) – The cancellation policy to use.

class zipline.finance.cancel_policy.CancelPolicy[source]#

Abstract cancellation policy interface.

abstract should_cancel(event)[source]#

Should all open orders be cancelled?

Parameters:

event (enum-value) –

An event type, one of:

zipline.gens.sim_engine.BAR
zipline.gens.sim_engine.DAY_START
zipline.gens.sim_engine.DAY_END
zipline.gens.sim_engine.MINUTE_END

Returns:

should_cancel – Should all open orders be cancelled?

Return type:

bool

zipline.api.EODCancel(warn_on_cancel=True)[source]#

This policy cancels open orders at the end of the day. For now, Zipline will only apply this policy to minutely simulations.

Parameters:: warn_on_cancel (bool, optional) – Should a warning be raised if this causes an order to be cancelled?

zipline.api.NeverCancel()[source]#: Orders are never automatically canceled.

Assets#

zipline.api.symbol(self, symbol_str, country_code=None)#

Lookup an Equity by its ticker symbol.

Parameters:

symbol_str (str) – The ticker symbol for the equity to lookup.
country_code (str or None, optional) – A country to limit symbol searches to.

Returns:

equity – The equity that held the ticker symbol on the current symbol lookup date.

Return type:

zipline.assets.Equity

Raises:

SymbolNotFound – Raised when the symbols was not held on the current lookup date.

See also

zipline.api.set_symbol_lookup_date()

zipline.api.symbols(self, *args, **kwargs)#

Lookup multuple Equities as a list.

Parameters:

*args (iterable[str]) – The ticker symbols to lookup.
country_code (str or None, optional) – A country to limit symbol searches to.

Returns:

equities – The equities that held the given ticker symbols on the current symbol lookup date.

Return type:

list[zipline.assets.Equity]

Raises:

SymbolNotFound – Raised when one of the symbols was not held on the current lookup date.

See also

zipline.api.set_symbol_lookup_date()

zipline.api.future_symbol(self, symbol)#

Lookup a futures contract with a given symbol.

Parameters:: symbol (str) – The symbol of the desired contract.
Returns:: future – The future that trades with the name symbol.
Return type:: zipline.assets.Future
Raises:: SymbolNotFound – Raised when no contract named ‘symbol’ is found.

zipline.api.set_symbol_lookup_date(self, dt)#

Set the date for which symbols will be resolved to their assets (symbols may map to different firms or underlying assets at different times)

Parameters:: dt (datetime) – The new symbol lookup date.

zipline.api.sid(self, sid)#

Lookup an Asset by its unique asset identifier.

Parameters:: sid (int) – The unique integer that identifies an asset.
Returns:: asset – The asset with the given sid.
Return type:: zipline.assets.Asset
Raises:: SidsNotFound – When a requested sid does not map to any asset.

Trading Controls#

Zipline provides trading controls to ensure that the algorithm performs as expected. The functions help protect the algorithm from undesirable consequences of unintended behavior, especially when trading with real money.

zipline.api.set_do_not_order_list(self, restricted_list, on_error='fail')#

Set a restriction on which assets can be ordered.

Parameters:: restricted_list (container[Asset], SecurityList) – The assets that cannot be ordered.

zipline.api.set_long_only(self, on_error='fail')#: Set a rule specifying that this algorithm cannot take short positions.

zipline.api.set_max_leverage(self, max_leverage)#

Set a limit on the maximum leverage of the algorithm.

Parameters:: max_leverage (float) – The maximum leverage for the algorithm. If not provided there will be no maximum.

zipline.api.set_max_order_count(self, max_count, on_error='fail')#

Set a limit on the number of orders that can be placed in a single day.

Parameters:: max_count (int) – The maximum number of orders that can be placed on any single day.

zipline.api.set_max_order_size(self, asset=None, max_shares=None, max_notional=None, on_error='fail')#

Set a limit on the number of shares and/or dollar value of any single order placed for sid. Limits are treated as absolute values and are enforced at the time that the algo attempts to place an order for sid.

If an algorithm attempts to place an order that would result in exceeding one of these limits, raise a TradingControlException.

Parameters:

asset (Asset, optional) – If provided, this sets the guard only on positions in the given asset.
max_shares (int, optional) – The maximum number of shares that can be ordered at one time.
max_notional (float, optional) – The maximum value that can be ordered at one time.

zipline.api.set_max_position_size(self, asset=None, max_shares=None, max_notional=None, on_error='fail')#

Set a limit on the number of shares and/or dollar value held for the given sid. Limits are treated as absolute values and are enforced at the time that the algo attempts to place an order for sid. This means that it’s possible to end up with more than the max number of shares due to splits/dividends, and more than the max notional due to price improvement.

If an algorithm attempts to place an order that would result in increasing the absolute value of shares/dollar value exceeding one of these limits, raise a TradingControlException.

Parameters:

asset (Asset, optional) – If provided, this sets the guard only on positions in the given asset.
max_shares (int, optional) – The maximum number of shares to hold for an asset.
max_notional (float, optional) – The maximum value to hold for an asset.

Simulation Parameters#

zipline.api.set_benchmark(self, benchmark)#

Set the benchmark asset.

Parameters:: benchmark (zipline.assets.Asset) – The asset to set as the new benchmark.

Notes

Any dividends payed out for that new benchmark asset will be automatically reinvested.

Commission Models#

zipline.api.set_commission(self, us_equities=None, us_futures=None)#

Sets the commission models for the simulation.

Parameters:

us_equities (EquityCommissionModel) – The commission model to use for trading US equities.
us_futures (FutureCommissionModel) – The commission model to use for trading US futures.

Notes

This function can only be called during initialize().

class zipline.finance.commission.CommissionModel[source]#

Abstract base class for commission models.

Commission models are responsible for accepting order/transaction pairs and calculating how much commission should be charged to an algorithm’s account on each transaction.

To implement a new commission model, create a subclass of CommissionModel and implement calculate().

abstract calculate(order, transaction)[source]#

Calculate the amount of commission to charge on order as a result of transaction.

Parameters:

order (zipline.finance.order.Order) –
The order being processed.

The commission field of order is a float indicating the amount of commission already charged on this order.
transaction (zipline.finance.transaction.Transaction) – The transaction being processed. A single order may generate multiple transactions if there isn’t enough volume in a given bar to fill the full amount requested in the order.

Returns:

amount_charged – The additional commission, in dollars, that we should attribute to this order.

Return type:

float

class zipline.finance.commission.PerShare(cost=0.001, min_trade_cost=0.0)[source]#

Calculates a commission for a transaction based on a per share cost with an optional minimum cost per trade.

Parameters:

cost (float, optional) – The amount of commissions paid per share traded. Default is one tenth of a cent per share.
min_trade_cost (float, optional) – The minimum amount of commissions paid per trade. Default is no minimum.

Notes

This is zipline’s default commission model for equities.

class zipline.finance.commission.PerTrade(cost=0.0)[source]#

Calculates a commission for a transaction based on a per trade cost.

For orders that require multiple fills, the full commission is charged to the first fill.

Parameters:: cost (float, optional) – The flat amount of commissions paid per equity trade.

class zipline.finance.commission.PerDollar(cost=0.0015)[source]#

Model commissions by applying a fixed cost per dollar transacted.

Parameters:: cost (float, optional) – The flat amount of commissions paid per dollar of equities traded. Default is a commission of $0.0015 per dollar transacted.

Slippage Models#

zipline.api.set_slippage(self, us_equities=None, us_futures=None)#

Set the slippage models for the simulation.

Parameters:

us_equities (EquitySlippageModel) – The slippage model to use for trading US equities.
us_futures (FutureSlippageModel) – The slippage model to use for trading US futures.

Notes

This function can only be called during initialize().

class zipline.finance.slippage.SlippageModel[source]#

Abstract base class for slippage models.

Slippage models are responsible for the rates and prices at which orders fill during a simulation.

To implement a new slippage model, create a subclass of SlippageModel and implement process_order().

process_order(data, order)[source]#

volume_for_bar#

Number of shares that have already been filled for the currently-filling asset in the current minute. This attribute is maintained automatically by the base class. It can be used by subclasses to keep track of the total amount filled if there are multiple open orders for a single asset.

Type:: int

Notes

Subclasses that define their own constructors should call super(<subclass name>, self).__init__() before performing other initialization.

abstract process_order(data, order)[source]#

Compute the number of shares and price to fill for order in the current minute.

Parameters:

data (zipline.protocol.BarData) – The data for the given bar.
order (zipline.finance.order.Order) – The order to simulate.

Returns:

execution_price (float) – The price of the fill.
execution_volume (int) – The number of shares that should be filled. Must be between 0 and order.amount - order.filled. If the amount filled is less than the amount remaining, order will remain open and will be passed again to this method in the next minute.

Raises:

zipline.finance.slippage.LiquidityExceeded – May be raised if no more orders should be processed for the current asset during the current bar.

Notes

Before this method is called, volume_for_bar will be set to the number of shares that have already been filled for order.asset in the current minute.

process_order() is not called by the base class on bars for which there was no historical volume.

class zipline.finance.slippage.FixedSlippage(spread=0.0)[source]#

Simple model assuming a fixed-size spread for all assets.

Parameters:: spread (float, optional) – Size of the assumed spread for all assets. Orders to buy will be filled at close + (spread / 2). Orders to sell will be filled at close - (spread / 2).

Notes

This model does not impose limits on the size of fills. An order for an asset will always be filled as soon as any trading activity occurs in the order’s asset, even if the size of the order is greater than the historical volume.

class zipline.finance.slippage.VolumeShareSlippage(volume_limit=0.025, price_impact=0.1)[source]#

Model slippage as a quadratic function of percentage of historical volume.

Orders to buy will be filled at:

price * (1 + price_impact * (volume_share ** 2))

Orders to sell will be filled at:

price * (1 - price_impact * (volume_share ** 2))

where price is the close price for the bar, and volume_share is the percentage of minutely volume filled, up to a max of volume_limit.

Parameters:

volume_limit (float, optional) – Maximum percent of historical volume that can fill in each bar. 0.5 means 50% of historical volume. 1.0 means 100%. Default is 0.025 (i.e., 2.5%).
price_impact (float, optional) – Scaling coefficient for price impact. Larger values will result in more simulated price impact. Smaller values will result in less simulated price impact. Default is 0.1.

Pipeline#

For more information, see Pipeline API

zipline.api.attach_pipeline(self, pipeline, name, chunks=None, eager=True)#

Parameters:

pipeline (Pipeline) – The pipeline to have computed.
name (str) – The name of the pipeline.
chunks (int or iterator, optional) – The number of days to compute pipeline results for. Increasing this number will make it longer to get the first results but may improve the total runtime of the simulation. If an iterator is passed, we will run in chunks based on values of the iterator. Default is True.
eager (bool, optional) – Whether or not to compute this pipeline prior to before_trading_start.

Returns:

pipeline – Returns the pipeline that was attached unchanged.

Return type:

Pipeline

See also

zipline.api.pipeline_output()

zipline.api.pipeline_output(self, name)#

Get results of the pipeline attached by with name name.

Parameters:: name (str) – Name of the pipeline from which to fetch results.
Returns:: results – DataFrame containing the results of the requested pipeline for the current simulation date.
Return type:: pd.DataFrame
Raises:: NoSuchPipeline – Raised when no pipeline with the name name has been registered.

Miscellaneous#

zipline.api.record(self, *args, **kwargs)#

Track and record values each day.

Parameters:: **kwargs – The names and values to record.

Notes

These values will appear in the performance packets and the performance dataframe passed to analyze and returned from run_algorithm().

zipline.api.get_environment(self, field='platform')#

Query the execution environment.

Parameters:

field ({'platform', 'arena', 'data_frequency', 'start', 'end',) –
'capital_base' –
'platform' –
'*'} –
meanings (The field to query. The options have the following) –
arena (-) – The arena from the simulation parameters. This will normally be 'backtest' but some systems may use this distinguish live trading from backtesting.
data_frequency (-) – data_frequency tells the algorithm if it is running with daily data or minute data.
start (-) – The start date for the simulation.
end (-) – The end date for the simulation.
capital_base (-) – The starting capital for the simulation.
-platform (str) – The platform that the code is running on. By default, this will be the string ‘zipline’. This can allow algorithms to know if they are running on the Quantopian platform instead.
* (-) – Returns all the fields in a dictionary.

Returns:

val – The value for the field queried. See above for more information.

Return type:

any

Raises:

ValueError – Raised when field is not a valid option.

zipline.api.fetch_csv(self, url, pre_func=None, post_func=None, date_column='date', date_format=None, timezone='UTC', symbol=None, mask=True, symbol_column=None, special_params_checker=None, country_code=None, **kwargs)#

Fetch a csv from a remote url and register the data so that it is queryable from the data object.

Parameters:

url (str) – The url of the csv file to load.
pre_func (callable[pd.DataFrame -> pd.DataFrame], optional) – A callback to allow preprocessing the raw data returned from fetch_csv before dates are paresed or symbols are mapped.
post_func (callable[pd.DataFrame -> pd.DataFrame], optional) – A callback to allow postprocessing of the data after dates and symbols have been mapped.
date_column (str, optional) – The name of the column in the preprocessed dataframe containing datetime information to map the data.
date_format (str, optional) – The format of the dates in the date_column. If not provided fetch_csv will attempt to infer the format. For information about the format of this string, see pandas.read_csv().
timezone (tzinfo or str, optional) – The timezone for the datetime in the date_column.
symbol (str, optional) – If the data is about a new asset or index then this string will be the name used to identify the values in data. For example, one may use fetch_csv to load data for VIX, then this field could be the string 'VIX'.
mask (bool, optional) – Drop any rows which cannot be symbol mapped.
symbol_column (str) – If the data is attaching some new attribute to each asset then this argument is the name of the column in the preprocessed dataframe containing the symbols. This will be used along with the date information to map the sids in the asset finder.
country_code (str, optional) – Country code to use to disambiguate symbol lookups.
**kwargs – Forwarded to pandas.read_csv().

Returns:

csv_data_source – A requests source that will pull data from the url specified.

Return type:

zipline.sources.requests_csv.PandasRequestsCSV

Blotters#

A blotter documents trades and their details over a period of time, typically one trading day. Trade details include such things as the time, price, order size, and whether it was a buy or sell order. It is is usually created by a trading software that records the trades made through a data feed.

class zipline.finance.blotter.blotter.Blotter(cancel_policy=None)[source]#

batch_order(order_arg_lists)[source]#

Place a batch of orders.

Parameters:: order_arg_lists (iterable[tuple]) – Tuples of args that order expects.
Returns:: order_ids – The unique identifier (or None) for each of the orders placed (or not placed).
Return type:: list[str or None]

Notes

This is required for Blotter subclasses to be able to place a batch of orders, instead of being passed the order requests one at a time.

abstract cancel(order_id, relay_status=True)[source]#

Cancel a single order

Parameters:

order_id (int) – The id of the order
relay_status (bool) – Whether or not to record the status of the order

abstract cancel_all_orders_for_asset(asset, warn=False, relay_status=True)[source]#: Cancel all open orders for a given asset.

abstract get_transactions(bar_data)[source]#

Creates a list of transactions based on the current open orders, slippage model, and commission model.

Parameters:: bar_data (zipline._protocol.BarData) –

Notes

This method book-keeps the blotter’s open_orders dictionary, so that: it is accurate by the time we’re done processing open orders.

Returns:

transactions_list (List) – transactions_list: list of transactions resulting from the current open orders. If there were no open orders, an empty list is returned.
commissions_list (List) – commissions_list: list of commissions resulting from filling the open orders. A commission is an object with “asset” and “cost” parameters.
closed_orders (List) – closed_orders: list of all the orders that have filled.

abstract hold(order_id, reason='')[source]#: Mark the order with order_id as ‘held’. Held is functionally similar to ‘open’. When a fill (full or partial) arrives, the status will automatically change back to open/filled as necessary.

abstract order(asset, amount, style, order_id=None)[source]#

Place an order.

Parameters:

asset (zipline.assets.Asset) – The asset that this order is for.
amount (int) – The amount of shares to order. If amount is positive, this is the number of shares to buy or cover. If amount is negative, this is the number of shares to sell or short.
style (zipline.finance.execution.ExecutionStyle) – The execution style for the order.
order_id (str, optional) – The unique identifier for this order.

Returns:

order_id – The unique identifier for this order, or None if no order was placed.

Return type:

str or None

Notes

amount > 0 : Buy/Cover amount < 0 : Sell/Short Market order : order(asset, amount) Limit order : order(asset, amount, style=LimitOrder(limit_price)) Stop order : order(asset, amount, style=StopOrder(stop_price)) StopLimit order : order(asset, amount, style=StopLimitOrder(limit_price, stop_price))

abstract process_splits(splits)[source]#

Processes a list of splits by modifying any open orders as needed.

Parameters:: splits (list) – A list of splits. Each split is a tuple of (asset, ratio).
Return type:: None

abstract prune_orders(closed_orders)[source]#

Removes all given orders from the blotter’s open_orders list.

Parameters:: closed_orders (iterable of orders that are closed.) –
Return type:: None

abstract reject(order_id, reason='')[source]#: Mark the given order as ‘rejected’, which is functionally similar to cancelled. The distinction is that rejections are involuntary (and usually include a message from a broker indicating why the order was rejected) while cancels are typically user-driven.

class zipline.finance.blotter.SimulationBlotter(equity_slippage=None, future_slippage=None, equity_commission=None, future_commission=None, cancel_policy=None)[source]#

cancel(order_id, relay_status=True)[source]#

Cancel a single order

Parameters:

order_id (int) – The id of the order
relay_status (bool) – Whether or not to record the status of the order

cancel_all_orders_for_asset(asset, warn=False, relay_status=True)[source]#: Cancel all open orders for a given asset.

get_transactions(bar_data)[source]#

Creates a list of transactions based on the current open orders, slippage model, and commission model.

Parameters:: bar_data (zipline._protocol.BarData) –

Notes

This method book-keeps the blotter’s open_orders dictionary, so that: it is accurate by the time we’re done processing open orders.

Returns:

transactions_list (List) – transactions_list: list of transactions resulting from the current open orders. If there were no open orders, an empty list is returned.
commissions_list (List) – commissions_list: list of commissions resulting from filling the open orders. A commission is an object with “asset” and “cost” parameters.
closed_orders (List) – closed_orders: list of all the orders that have filled.

hold(order_id, reason='')[source]#: Mark the order with order_id as ‘held’. Held is functionally similar to ‘open’. When a fill (full or partial) arrives, the status will automatically change back to open/filled as necessary.

order(asset, amount, style, order_id=None)[source]#

Place an order.

Parameters:

asset (zipline.assets.Asset) – The asset that this order is for.
amount (int) – The amount of shares to order. If amount is positive, this is the number of shares to buy or cover. If amount is negative, this is the number of shares to sell or short.
style (zipline.finance.execution.ExecutionStyle) – The execution style for the order.
order_id (str, optional) – The unique identifier for this order.

Returns:

order_id – The unique identifier for this order, or None if no order was placed.

Return type:

str or None

Notes

amount > 0 :: Buy/Cover amount < 0 :: Sell/Short Market order: order(asset, amount) Limit order: order(asset, amount, style=LimitOrder(limit_price)) Stop order: order(asset, amount, style=StopOrder(stop_price)) StopLimit order: order(asset, amount, style=StopLimitOrder(limit_price, stop_price))

process_splits(splits)[source]#

Processes a list of splits by modifying any open orders as needed.

Parameters:: splits (list) – A list of splits. Each split is a tuple of (asset, ratio).
Return type:: None

prune_orders(closed_orders)[source]#

Removes all given orders from the blotter’s open_orders list.

Parameters:: closed_orders (iterable of orders that are closed.) –
Return type:: None

reject(order_id, reason='')[source]#: Mark the given order as ‘rejected’, which is functionally similar to cancelled. The distinction is that rejections are involuntary (and usually include a message from a broker indicating why the order was rejected) while cancels are typically user-driven.

Pipeline API#

A Pipeline enables faster and more memory-efficient execution by optimizing the computation of factors during a backtest.

class zipline.pipeline.Pipeline(columns=None, screen=None, domain=GENERIC)[source]#

A Pipeline object represents a collection of named expressions to be compiled and executed by a PipelineEngine.

A Pipeline has two important attributes: ‘columns’, a dictionary of named Term instances, and ‘screen’, a Filter representing criteria for including an asset in the results of a Pipeline.

To compute a pipeline in the context of a TradingAlgorithm, users must call attach_pipeline in their initialize function to register that the pipeline should be computed each trading day. The most recent outputs of an attached pipeline can be retrieved by calling pipeline_output from handle_data, before_trading_start, or a scheduled function.

Parameters:

columns (dict, optional) – Initial columns.
screen (zipline.pipeline.Filter, optional) – Initial screen.

add(term, name, overwrite=False)[source]#

Add a column.

The results of computing term will show up as a column in the DataFrame produced by running this pipeline.

Parameters:

column (zipline.pipeline.Term) – A Filter, Factor, or Classifier to add to the pipeline.
name (str) – Name of the column to add.
overwrite (bool) – Whether to overwrite the existing entry if we already have a column named name.

domain(default)[source]#

Get the domain for this pipeline.

If an explicit domain was provided at construction time, use it.
Otherwise, infer a domain from the registered columns.
If no domain can be inferred, return default.

Parameters:

default (zipline.pipeline.domain.Domain) – Domain to use if no domain can be inferred from this pipeline by itself.

Returns:

domain – The domain for the pipeline.

Return type:

zipline.pipeline.domain.Domain

Raises:

AmbiguousDomain –
ValueError – If the terms in self conflict with self._domain.

remove(name)[source]#

Remove a column.

Parameters:: name (str) – The name of the column to remove.
Raises:: KeyError – If name is not in self.columns.
Returns:: removed – The removed term.
Return type:: zipline.pipeline.Term

set_screen(screen, overwrite=False)[source]#

Set a screen on this Pipeline.

Parameters:

filter (zipline.pipeline.Filter) – The filter to apply as a screen.
overwrite (bool) – Whether to overwrite any existing screen. If overwrite is False and self.screen is not None, we raise an error.

show_graph(format='svg')[source]#

Render this Pipeline as a DAG.

Parameters:: format ({'svg', 'png', 'jpeg'}) – Image format to render with. Default is ‘svg’.

to_execution_plan(domain, default_screen, start_date, end_date)[source]#

Compile into an ExecutionPlan.

Parameters:

domain (zipline.pipeline.domain.Domain) – Domain on which the pipeline will be executed.
default_screen (zipline.pipeline.Term) – Term to use as a screen if self.screen is None.
all_dates (pd.DatetimeIndex) – A calendar of dates to use to calculate starts and ends for each term.
start_date (pd.Timestamp) – The first date of requested output.
end_date (pd.Timestamp) – The last date of requested output.

Returns:

graph – Graph encoding term dependencies, including metadata about extra row requirements.

Return type:

zipline.pipeline.graph.ExecutionPlan

to_simple_graph(default_screen)[source]#

Compile into a simple TermGraph with no extra row metadata.

Parameters:: default_screen (zipline.pipeline.Term) – Term to use as a screen if self.screen is None.
Returns:: graph – Graph encoding term dependencies.
Return type:: zipline.pipeline.graph.TermGraph

property columns#

The output columns of this pipeline.

Returns:: columns – Map from column name to expression computing that column’s output.
Return type:: dict[str, zipline.pipeline.ComputableTerm]

property screen#

The screen of this pipeline.

Returns:: screen – Term defining the screen for this pipeline. If screen is a filter, rows that do not pass the filter (i.e., rows for which the filter computed False) will be dropped from the output of this pipeline before returning results.
Return type:: zipline.pipeline.Filter or None

Notes

Setting a screen on a Pipeline does not change the values produced for any rows: it only affects whether a given row is returned. Computing a pipeline with a screen is logically equivalent to computing the pipeline without the screen and then, as a post-processing-step, filtering out any rows for which the screen computed False.

class zipline.pipeline.CustomFactor(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Base class for user-defined Factors.

Parameters:

inputs (iterable, optional) – An iterable of BoundColumn instances (e.g. USEquityPricing.close), describing the data to load and pass to self.compute. If this argument is not passed to the CustomFactor constructor, we look for a class-level attribute named inputs.
outputs (iterable[str], optional) – An iterable of strings which represent the names of each output this factor should compute and return. If this argument is not passed to the CustomFactor constructor, we look for a class-level attribute named outputs.
window_length (int, optional) – Number of rows to pass for each input. If this argument is not passed to the CustomFactor constructor, we look for a class-level attribute named window_length.
mask (zipline.pipeline.Filter, optional) – A Filter describing the assets on which we should compute each day. Each call to CustomFactor.compute will only receive assets for which mask produced True on the day for which compute is being called.

Notes

Users implementing their own Factors should subclass CustomFactor and implement a method named compute with the following signature:

def compute(self, today, assets, out, *inputs):
   ...

On each simulation date, compute will be called with the current date, an array of sids, an output array, and an input array for each expression passed as inputs to the CustomFactor constructor.

The specific types of the values passed to compute are as follows:

today : np.datetime64[ns]
    Row label for the last row of all arrays passed as `inputs`.
assets : np.array[int64, ndim=1]
    Column labels for `out` and`inputs`.
out : np.array[self.dtype, ndim=1]
    Output array of the same shape as `assets`.  `compute` should write
    its desired return values into `out`. If multiple outputs are
    specified, `compute` should write its desired return values into
    `out.<output_name>` for each output name in `self.outputs`.
*inputs : tuple of np.array
    Raw data arrays corresponding to the values of `self.inputs`.

compute functions should expect to be passed NaN values for dates on which no data was available for an asset. This may include dates on which an asset did not yet exist.

For example, if a CustomFactor requires 10 rows of close price data, and asset A started trading on Monday June 2nd, 2014, then on Tuesday, June 3rd, 2014, the column of input data for asset A will have 9 leading NaNs for the preceding days on which data was not yet available.

Examples

A CustomFactor with pre-declared defaults:

class TenDayRange(CustomFactor):
    """
    Computes the difference between the highest high in the last 10
    days and the lowest low.

    Pre-declares high and low as default inputs and `window_length` as
    10.
    """

    inputs = [USEquityPricing.high, USEquityPricing.low]
    window_length = 10

    def compute(self, today, assets, out, highs, lows):
        from numpy import nanmin, nanmax

        highest_highs = nanmax(highs, axis=0)
        lowest_lows = nanmin(lows, axis=0)
        out[:] = highest_highs - lowest_lows

# Doesn't require passing inputs or window_length because they're
# pre-declared as defaults for the TenDayRange class.
ten_day_range = TenDayRange()

A CustomFactor without defaults:

class MedianValue(CustomFactor):
    """
    Computes the median value of an arbitrary single input over an
    arbitrary window..

    Does not declare any defaults, so values for `window_length` and
    `inputs` must be passed explicitly on every construction.
    """

    def compute(self, today, assets, out, data):
        from numpy import nanmedian
        out[:] = data.nanmedian(data, axis=0)

# Values for `inputs` and `window_length` must be passed explicitly to
# MedianValue.
median_close10 = MedianValue([USEquityPricing.close], window_length=10)
median_low15 = MedianValue([USEquityPricing.low], window_length=15)

A CustomFactor with multiple outputs:

class MultipleOutputs(CustomFactor):
    inputs = [USEquityPricing.close]
    outputs = ['alpha', 'beta']
    window_length = N

    def compute(self, today, assets, out, close):
        computed_alpha, computed_beta = some_function(close)
        out.alpha[:] = computed_alpha
        out.beta[:] = computed_beta

# Each output is returned as its own Factor upon instantiation.
alpha, beta = MultipleOutputs()

# Equivalently, we can create a single factor instance and access each
# output as an attribute of that instance.
multiple_outputs = MultipleOutputs()
alpha = multiple_outputs.alpha
beta = multiple_outputs.beta

Note: If a CustomFactor has multiple outputs, all outputs must have the same dtype. For instance, in the example above, if alpha is a float then beta must also be a float.

dtype = dtype('float64')#

class zipline.pipeline.Filter(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), domain=sentinel('NotSpecified'), *args, **kwargs)[source]#

Pipeline expression computing a boolean output.

Filters are most commonly useful for describing sets of assets to include or exclude for some particular purpose. Many Pipeline API functions accept a mask argument, which can be supplied a Filter indicating that only values passing the Filter should be considered when performing the requested computation. For example, zipline.pipeline.Factor.top() accepts a mask indicating that ranks should be computed only on assets that passed the specified Filter.

The most common way to construct a Filter is via one of the comparison operators (<, <=, !=, eq, >, >=) of Factor. For example, a natural way to construct a Filter for stocks with a 10-day VWAP less than $20.0 is to first construct a Factor computing 10-day VWAP and compare it to the scalar value 20.0:

>>> from zipline.pipeline.factors import VWAP
>>> vwap_10 = VWAP(window_length=10)
>>> vwaps_under_20 = (vwap_10 <= 20)

Filters can also be constructed via comparisons between two Factors. For example, to construct a Filter producing True for asset/date pairs where the asset’s 10-day VWAP was greater than it’s 30-day VWAP:

>>> short_vwap = VWAP(window_length=10)
>>> long_vwap = VWAP(window_length=30)
>>> higher_short_vwap = (short_vwap > long_vwap)

Filters can be combined via the & (and) and | (or) operators.

&-ing together two filters produces a new Filter that produces True if both of the inputs produced True.

|-ing together two filters produces a new Filter that produces True if either of its inputs produced True.

The ~ operator can be used to invert a Filter, swapping all True values with Falses and vice-versa.

Filters may be set as the screen attribute of a Pipeline, indicating asset/date pairs for which the filter produces False should be excluded from the Pipeline’s output. This is useful both for reducing noise in the output of a Pipeline and for reducing memory consumption of Pipeline results.

__and__(other)#: Binary Operator: ‘&’

__or__(other)#: Binary Operator: ‘|’

if_else(if_true, if_false)[source]#

Create a term that selects values from one of two choices.

Parameters:

if_true (zipline.pipeline.term.ComputableTerm) – Expression whose values should be used at locations where this filter outputs True.
if_false (zipline.pipeline.term.ComputableTerm) – Expression whose values should be used at locations where this filter outputs False.

Returns:

merged – A term that computes by taking values from either if_true or if_false, depending on the values produced by self.

The returned term draws from``if_true`` at locations where self produces True, and it draws from if_false at locations where self produces False.

Return type:

zipline.pipeline.term.ComputableTerm

Example

Let f be a Factor that produces the following output:

             AAPL   MSFT    MCD     BK
2017-03-13    1.0    2.0    3.0    4.0
2017-03-14    5.0    6.0    7.0    8.0

Let g be another Factor that produces the following output:

             AAPL   MSFT    MCD     BK
2017-03-13   10.0   20.0   30.0   40.0
2017-03-14   50.0   60.0   70.0   80.0

Finally, let condition be a Filter that produces the following output:

             AAPL   MSFT    MCD     BK
2017-03-13   True  False   True  False
2017-03-14   True   True  False  False

Then, the expression condition.if_else(f, g) produces the following output:

             AAPL   MSFT    MCD     BK
2017-03-13    1.0   20.0    3.0   40.0
2017-03-14    5.0    6.0   70.0   80.0

See also

numpy.where, Factor.fillna

class zipline.pipeline.Factor(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), domain=sentinel('NotSpecified'), *args, **kwargs)[source]#

Pipeline API expression producing a numerical or date-valued output.

Factors are the most commonly-used Pipeline term, representing the result of any computation producing a numerical result.

Factors can be combined, both with other Factors and with scalar values, via any of the builtin mathematical operators (+, -, *, etc).

This makes it easy to write complex expressions that combine multiple Factors. For example, constructing a Factor that computes the average of two other Factors is simply:

>>> f1 = SomeFactor(...)  
>>> f2 = SomeOtherFactor(...)  
>>> average = (f1 + f2) / 2.0  

Factors can also be converted into zipline.pipeline.Filter objects via comparison operators: (<, <=, !=, eq, >, >=).

There are many natural operators defined on Factors besides the basic numerical operators. These include methods for identifying missing or extreme-valued outputs (isnull(), notnull(), isnan(), notnan()), methods for normalizing outputs (rank(), demean(), zscore()), and methods for constructing Filters based on rank-order properties of results (top(), bottom(), percentile_between()).

eq(other)#

Construct a Filter computing self == other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: filter – Filter computing self == other with the outputs of self and other.
Return type:: zipline.pipeline.Filter

demean(mask=sentinel('NotSpecified'), groupby=sentinel('NotSpecified'))[source]#

Construct a Factor that computes self and subtracts the mean from row of the result.

If mask is supplied, ignore values where mask returns False when computing row means, and output NaN anywhere the mask is False.

If groupby is supplied, compute by partitioning each row based on the values produced by groupby, de-meaning the partitioned arrays, and stitching the sub-results back together.

Parameters:

mask (zipline.pipeline.Filter, optional) – A Filter defining values to ignore when computing means.
groupby (zipline.pipeline.Classifier, optional) – A classifier defining partitions over which to compute means.

Examples

Let f be a Factor which would produce the following output:

             AAPL   MSFT    MCD     BK
2017-03-13    1.0    2.0    3.0    4.0
2017-03-14    1.5    2.5    3.5    1.0
2017-03-15    2.0    3.0    4.0    1.5
2017-03-16    2.5    3.5    1.0    2.0

Let c be a Classifier producing the following output:

             AAPL   MSFT    MCD     BK
2017-03-13      1      1      2      2
2017-03-14      1      1      2      2
2017-03-15      1      1      2      2
2017-03-16      1      1      2      2

Let m be a Filter producing the following output:

             AAPL   MSFT    MCD     BK
2017-03-13  False   True   True   True
2017-03-14   True  False   True   True
2017-03-15   True   True  False   True
2017-03-16   True   True   True  False

Then f.demean() will subtract the mean from each row produced by f.

             AAPL   MSFT    MCD     BK
2017-03-13 -1.500 -0.500  0.500  1.500
2017-03-14 -0.625  0.375  1.375 -1.125
2017-03-15 -0.625  0.375  1.375 -1.125
2017-03-16  0.250  1.250 -1.250 -0.250

f.demean(mask=m) will subtract the mean from each row, but means will be calculated ignoring values on the diagonal, and NaNs will written to the diagonal in the output. Diagonal values are ignored because they are the locations where the mask m produced False.

             AAPL   MSFT    MCD     BK
2017-03-13    NaN -1.000  0.000  1.000
2017-03-14 -0.500    NaN  1.500 -1.000
2017-03-15 -0.166  0.833    NaN -0.666
2017-03-16  0.166  1.166 -1.333    NaN

f.demean(groupby=c) will subtract the group-mean of AAPL/MSFT and MCD/BK from their respective entries. The AAPL/MSFT are grouped together because both assets always produce 1 in the output of the classifier c. Similarly, MCD/BK are grouped together because they always produce 2.

             AAPL   MSFT    MCD     BK
2017-03-13 -0.500  0.500 -0.500  0.500
2017-03-14 -0.500  0.500  1.250 -1.250
2017-03-15 -0.500  0.500  1.250 -1.250
2017-03-16 -0.500  0.500 -0.500  0.500

f.demean(mask=m, groupby=c) will also subtract the group-mean of AAPL/MSFT and MCD/BK, but means will be calculated ignoring values on the diagonal , and NaNs will be written to the diagonal in the output.

             AAPL   MSFT    MCD     BK
2017-03-13    NaN  0.000 -0.500  0.500
2017-03-14  0.000    NaN  1.250 -1.250
2017-03-15 -0.500  0.500    NaN  0.000
2017-03-16 -0.500  0.500  0.000    NaN

Notes

Mean is sensitive to the magnitudes of outliers. When working with factor that can potentially produce large outliers, it is often useful to use the mask parameter to discard values at the extremes of the distribution:

>>> base = MyFactor(...)  
>>> normalized = base.demean(
...     mask=base.percentile_between(1, 99),
... )  

demean() is only supported on Factors of dtype float64.

See also

pandas.DataFrame.groupby()

zscore(mask=sentinel('NotSpecified'), groupby=sentinel('NotSpecified'))[source]#

Construct a Factor that Z-Scores each day’s results.

The Z-Score of a row is defined as:

(row - row.mean()) / row.stddev()

If mask is supplied, ignore values where mask returns False when computing row means and standard deviations, and output NaN anywhere the mask is False.

If groupby is supplied, compute by partitioning each row based on the values produced by groupby, z-scoring the partitioned arrays, and stitching the sub-results back together.

Parameters:

mask (zipline.pipeline.Filter, optional) – A Filter defining values to ignore when Z-Scoring.
groupby (zipline.pipeline.Classifier, optional) – A classifier defining partitions over which to compute Z-Scores.

Returns:

zscored – A Factor producing that z-scores the output of self.

Return type:

zipline.pipeline.Factor

Notes

Mean and standard deviation are sensitive to the magnitudes of outliers. When working with factor that can potentially produce large outliers, it is often useful to use the mask parameter to discard values at the extremes of the distribution:

>>> base = MyFactor(...)  
>>> normalized = base.zscore(
...    mask=base.percentile_between(1, 99),
... )  

zscore() is only supported on Factors of dtype float64.

Examples

See demean() for an in-depth example of the semantics for mask and groupby.

See also

pandas.DataFrame.groupby()

rank(method='ordinal', ascending=True, mask=sentinel('NotSpecified'), groupby=sentinel('NotSpecified'))[source]#

Construct a new Factor representing the sorted rank of each column within each row.

Parameters:

method (str, {'ordinal', 'min', 'max', 'dense', 'average'}) – The method used to assign ranks to tied elements. See scipy.stats.rankdata for a full description of the semantics for each ranking method. Default is ‘ordinal’.
ascending (bool, optional) – Whether to return sorted rank in ascending or descending order. Default is True.
mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when computing ranks. If mask is supplied, ranks are computed ignoring any asset/date pairs for which mask produces a value of False.
groupby (zipline.pipeline.Classifier, optional) – A classifier defining partitions over which to perform ranking.

Returns:

ranks – A new factor that will compute the ranking of the data produced by self.

Return type:

zipline.pipeline.Factor

Notes

The default value for method is different from the default for scipy.stats.rankdata. See that function’s documentation for a full description of the valid inputs to method.

Missing or non-existent data on a given day will cause an asset to be given a rank of NaN for that day.

See also

scipy.stats.rankdata()

pearsonr(target, correlation_length, mask=sentinel('NotSpecified'))[source]#

Construct a new Factor that computes rolling pearson correlation coefficients between target and the columns of self.

Parameters:

target (zipline.pipeline.Term) – The term used to compute correlations against each column of data produced by self. This may be a Factor, a BoundColumn or a Slice. If target is two-dimensional, correlations are computed asset-wise.
correlation_length (int) – Length of the lookback window over which to compute each correlation coefficient.
mask (zipline.pipeline.Filter, optional) – A Filter describing which assets should have their correlation with the target slice computed each day.

Returns:

correlations – A new Factor that will compute correlations between target and the columns of self.

Return type:

zipline.pipeline.Factor

Notes

This method can only be called on expressions which are deemed safe for use as inputs to windowed Factor objects. Examples of such expressions include This includes BoundColumn Returns and any factors created from rank() or zscore().

Examples

Suppose we want to create a factor that computes the correlation between AAPL’s 10-day returns and the 10-day returns of all other assets, computing each correlation over 30 days. This can be achieved by doing the following:

returns = Returns(window_length=10)
returns_slice = returns[sid(24)]
aapl_correlations = returns.pearsonr(
    target=returns_slice, correlation_length=30,
)

This is equivalent to doing:

aapl_correlations = RollingPearsonOfReturns(
    target=sid(24), returns_length=10, correlation_length=30,
)

spearmanr(target, correlation_length, mask=sentinel('NotSpecified'))[source]#

Construct a new Factor that computes rolling spearman rank correlation coefficients between target and the columns of self.

Parameters:

target (zipline.pipeline.Term) – The term used to compute correlations against each column of data produced by self. This may be a Factor, a BoundColumn or a Slice. If target is two-dimensional, correlations are computed asset-wise.
correlation_length (int) – Length of the lookback window over which to compute each correlation coefficient.
mask (zipline.pipeline.Filter, optional) – A Filter describing which assets should have their correlation with the target slice computed each day.

Returns:

correlations – A new Factor that will compute correlations between target and the columns of self.

Return type:

zipline.pipeline.Factor

Notes

Examples

returns = Returns(window_length=10)
returns_slice = returns[sid(24)]
aapl_correlations = returns.spearmanr(
    target=returns_slice, correlation_length=30,
)

This is equivalent to doing:

aapl_correlations = RollingSpearmanOfReturns(
    target=sid(24), returns_length=10, correlation_length=30,
)

linear_regression(target, regression_length, mask=sentinel('NotSpecified'))[source]#

Construct a new Factor that performs an ordinary least-squares regression predicting the columns of self from target.

Parameters:

target (zipline.pipeline.Term) – The term to use as the predictor/independent variable in each regression. This may be a Factor, a BoundColumn or a Slice. If target is two-dimensional, regressions are computed asset-wise.
regression_length (int) – Length of the lookback window over which to compute each regression.
mask (zipline.pipeline.Filter, optional) – A Filter describing which assets should be regressed with the target slice each day.

Returns:

regressions – A new Factor that will compute linear regressions of target against the columns of self.

Return type:

zipline.pipeline.Factor

Notes

Examples

Suppose we want to create a factor that regresses AAPL’s 10-day returns against the 10-day returns of all other assets, computing each regression over 30 days. This can be achieved by doing the following:

returns = Returns(window_length=10)
returns_slice = returns[sid(24)]
aapl_regressions = returns.linear_regression(
    target=returns_slice, regression_length=30,
)

This is equivalent to doing:

aapl_regressions = RollingLinearRegressionOfReturns(
    target=sid(24), returns_length=10, regression_length=30,
)

See also

scipy.stats.linregress()

winsorize(min_percentile, max_percentile, mask=sentinel('NotSpecified'), groupby=sentinel('NotSpecified'))[source]#

Construct a new factor that winsorizes the result of this factor.

Winsorizing changes values ranked less than the minimum percentile to the value at the minimum percentile. Similarly, values ranking above the maximum percentile are changed to the value at the maximum percentile.

Winsorizing is useful for limiting the impact of extreme data points without completely removing those points.

If mask is supplied, ignore values where mask returns False when computing percentile cutoffs, and output NaN anywhere the mask is False.

If groupby is supplied, winsorization is applied separately separately to each group defined by groupby.

Parameters:

min_percentile (float, int) – Entries with values at or below this percentile will be replaced with the (len(input) * min_percentile)th lowest value. If low values should not be clipped, use 0.
max_percentile (float, int) – Entries with values at or above this percentile will be replaced with the (len(input) * max_percentile)th lowest value. If high values should not be clipped, use 1.
mask (zipline.pipeline.Filter, optional) – A Filter defining values to ignore when winsorizing.
groupby (zipline.pipeline.Classifier, optional) – A classifier defining partitions over which to winsorize.

Returns:

winsorized – A Factor producing a winsorized version of self.

Return type:

zipline.pipeline.Factor

Examples

price = USEquityPricing.close.latest
columns={
    'PRICE': price,
    'WINSOR_1: price.winsorize(
        min_percentile=0.25, max_percentile=0.75
    ),
    'WINSOR_2': price.winsorize(
        min_percentile=0.50, max_percentile=1.0
    ),
    'WINSOR_3': price.winsorize(
        min_percentile=0.0, max_percentile=0.5
    ),

}

Given a pipeline with columns, defined above, the result for a given day could look like:

        'PRICE' 'WINSOR_1' 'WINSOR_2' 'WINSOR_3'
Asset_1    1        2          4          3
Asset_2    2        2          4          3
Asset_3    3        3          4          3
Asset_4    4        4          4          4
Asset_5    5        5          5          4
Asset_6    6        5          5          4

quantiles(bins, mask=sentinel('NotSpecified'))[source]#

Construct a Classifier computing quantiles of the output of self.

Every non-NaN data point the output is labelled with an integer value from 0 to (bins - 1). NaNs are labelled with -1.

If mask is supplied, ignore data points in locations for which mask produces False, and emit a label of -1 at those locations.

Parameters:

bins (int) – Number of bins labels to compute.
mask (zipline.pipeline.Filter, optional) – Mask of values to ignore when computing quantiles.

Returns:

quantiles – A classifier producing integer labels ranging from 0 to (bins - 1).

Return type:

zipline.pipeline.Classifier

quartiles(mask=sentinel('NotSpecified'))[source]#

Construct a Classifier computing quartiles over the output of self.

Every non-NaN data point the output is labelled with a value of either 0, 1, 2, or 3, corresponding to the first, second, third, or fourth quartile over each row. NaN data points are labelled with -1.

If mask is supplied, ignore data points in locations for which mask produces False, and emit a label of -1 at those locations.

Parameters:: mask (zipline.pipeline.Filter, optional) – Mask of values to ignore when computing quartiles.
Returns:: quartiles – A classifier producing integer labels ranging from 0 to 3.
Return type:: zipline.pipeline.Classifier

quintiles(mask=sentinel('NotSpecified'))[source]#

Construct a Classifier computing quintile labels on self.

Every non-NaN data point the output is labelled with a value of either 0, 1, 2, or 3, 4, corresonding to quintiles over each row. NaN data points are labelled with -1.

If mask is supplied, ignore data points in locations for which mask produces False, and emit a label of -1 at those locations.

Parameters:: mask (zipline.pipeline.Filter, optional) – Mask of values to ignore when computing quintiles.
Returns:: quintiles – A classifier producing integer labels ranging from 0 to 4.
Return type:: zipline.pipeline.Classifier

deciles(mask=sentinel('NotSpecified'))[source]#

Construct a Classifier computing decile labels on self.

Every non-NaN data point the output is labelled with a value from 0 to 9 corresonding to deciles over each row. NaN data points are labelled with -1.

If mask is supplied, ignore data points in locations for which mask produces False, and emit a label of -1 at those locations.

Parameters:: mask (zipline.pipeline.Filter, optional) – Mask of values to ignore when computing deciles.
Returns:: deciles – A classifier producing integer labels ranging from 0 to 9.
Return type:: zipline.pipeline.Classifier

top(N, mask=sentinel('NotSpecified'), groupby=sentinel('NotSpecified'))[source]#

Construct a Filter matching the top N asset values of self each day.

If groupby is supplied, returns a Filter matching the top N asset values for each group.

Parameters:

N (int) – Number of assets passing the returned filter each day.
mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when computing ranks. If mask is supplied, top values are computed ignoring any asset/date pairs for which mask produces a value of False.
groupby (zipline.pipeline.Classifier, optional) – A classifier defining partitions over which to perform ranking.

Returns:

filter

Return type:

zipline.pipeline.Filter

bottom(N, mask=sentinel('NotSpecified'), groupby=sentinel('NotSpecified'))[source]#

Construct a Filter matching the bottom N asset values of self each day.

If groupby is supplied, returns a Filter matching the bottom N asset values for each group defined by groupby.

Parameters:

N (int) – Number of assets passing the returned filter each day.
mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when computing ranks. If mask is supplied, bottom values are computed ignoring any asset/date pairs for which mask produces a value of False.
groupby (zipline.pipeline.Classifier, optional) – A classifier defining partitions over which to perform ranking.

Returns:

filter

Return type:

zipline.pipeline.Filter

percentile_between(min_percentile, max_percentile, mask=sentinel('NotSpecified'))[source]#

Construct a Filter matching values of self that fall within the range defined by min_percentile and max_percentile.

Parameters:

min_percentile (float [0.0, 100.0]) – Return True for assets falling above this percentile in the data.
max_percentile (float [0.0, 100.0]) – Return True for assets falling below this percentile in the data.
mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when percentile calculating thresholds. If mask is supplied, percentile cutoffs are computed each day using only assets for which mask returns True. Assets for which mask produces False will produce False in the output of this Factor as well.

Returns:

out – A new filter that will compute the specified percentile-range mask.

Return type:

zipline.pipeline.Filter

isnan()[source]#

A Filter producing True for all values where this Factor is NaN.

Returns:: nanfilter
Return type:: zipline.pipeline.Filter

notnan()[source]#

A Filter producing True for values where this Factor is not NaN.

Returns:: nanfilter
Return type:: zipline.pipeline.Filter

isfinite()[source]#: A Filter producing True for values where this Factor is anything but NaN, inf, or -inf.

clip(min_bound, max_bound, mask=sentinel('NotSpecified'))[source]#

Clip (limit) the values in a factor.

Given an interval, values outside the interval are clipped to the interval edges. For example, if an interval of [0, 1] is specified, values smaller than 0 become 0, and values larger than 1 become 1.

Parameters:

min_bound (float) – The minimum value to use.
max_bound (float) – The maximum value to use.
mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when clipping.

Notes

To only clip values on one side, -np.inf` and ``np.inf may be passed. For example, to only clip the maximum value but not clip a minimum value:

factor.clip(min_bound=-np.inf, max_bound=user_provided_max)

See also

numpy.clip

clip(min_bound, max_bound, mask=sentinel('NotSpecified'))[source]#

Clip (limit) the values in a factor.

Parameters:

min_bound (float) – The minimum value to use.
max_bound (float) – The maximum value to use.
mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when clipping.

Notes

To only clip values on one side, -np.inf` and ``np.inf may be passed. For example, to only clip the maximum value but not clip a minimum value:

factor.clip(min_bound=-np.inf, max_bound=user_provided_max)

See also

numpy.clip

__add__(other)#

Construct a Factor computing self + other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: factor – Factor computing self + other with outputs of self and other.
Return type:: zipline.pipeline.Factor

__sub__(other)#

Construct a Factor computing self - other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: factor – Factor computing self - other with outputs of self and other.
Return type:: zipline.pipeline.Factor

__mul__(other)#

Construct a Factor computing self * other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: factor – Factor computing self * other with outputs of self and other.
Return type:: zipline.pipeline.Factor

__div__(other)#

Construct a Factor computing self / other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: factor – Factor computing self / other with outputs of self and other.
Return type:: zipline.pipeline.Factor

__mod__(other)#

Construct a Factor computing self % other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: factor – Factor computing self % other with outputs of self and other.
Return type:: zipline.pipeline.Factor

__pow__(other)#

Construct a Factor computing self ** other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: factor – Factor computing self ** other with outputs of self and other.
Return type:: zipline.pipeline.Factor

__lt__(other)#

Construct a Filter computing self < other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: filter – Filter computing self < other with the outputs of self and other.
Return type:: zipline.pipeline.Filter

__le__(other)#

Construct a Filter computing self <= other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: filter – Filter computing self <= other with the outputs of self and other.
Return type:: zipline.pipeline.Filter

__ne__(other)#

Construct a Filter computing self != other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: filter – Filter computing self != other with the outputs of self and other.
Return type:: zipline.pipeline.Filter

__ge__(other)#

Construct a Filter computing self >= other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: filter – Filter computing self >= other with the outputs of self and other.
Return type:: zipline.pipeline.Filter

__gt__(other)#

Construct a Filter computing self > other.

Parameters:: other (zipline.pipeline.Factor, float) – Right-hand side of the expression.
Returns:: filter – Filter computing self > other with the outputs of self and other.
Return type:: zipline.pipeline.Filter

fillna(fill_value)#

Create a new term that fills missing values of this term’s output with fill_value.

Parameters:

fill_value (zipline.pipeline.ComputableTerm, or object.) –

Object to use as replacement for missing values.

If a ComputableTerm (e.g. a Factor) is passed, that term’s results will be used as fill values.

If a scalar (e.g. a number) is passed, the scalar will be used as a fill value.

Examples

Filling with a Scalar:

Let f be a Factor which would produce the following output:

             AAPL   MSFT    MCD     BK
2017-03-13    1.0    NaN    3.0    4.0
2017-03-14    1.5    2.5    NaN    NaN

Then f.fillna(0) produces the following output:

             AAPL   MSFT    MCD     BK
2017-03-13    1.0    0.0    3.0    4.0
2017-03-14    1.5    2.5    0.0    0.0

Filling with a Term:

Let f be as above, and let g be another Factor which would produce the following output:

             AAPL   MSFT    MCD     BK
2017-03-13   10.0   20.0   30.0   40.0
2017-03-14   15.0   25.0   35.0   45.0

Then, f.fillna(g) produces the following output:

             AAPL   MSFT    MCD     BK
2017-03-13    1.0   20.0    3.0    4.0
2017-03-14    1.5    2.5   35.0   45.0

Returns:: filled – A term computing the same results as self, but with missing values filled in using values from fill_value.
Return type:: zipline.pipeline.ComputableTerm

mean(mask=sentinel('NotSpecified'))#

Create a 1-dimensional factor computing the mean of self, each day.

Parameters:: mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when computing results. If supplied, we ignore asset/date pairs where mask produces False.
Returns:: result
Return type:: zipline.pipeline.Factor

stddev(mask=sentinel('NotSpecified'))#

Create a 1-dimensional factor computing the stddev of self, each day.

Parameters:: mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when computing results. If supplied, we ignore asset/date pairs where mask produces False.
Returns:: result
Return type:: zipline.pipeline.Factor

max(mask=sentinel('NotSpecified'))#

Create a 1-dimensional factor computing the max of self, each day.

Parameters:: mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when computing results. If supplied, we ignore asset/date pairs where mask produces False.
Returns:: result
Return type:: zipline.pipeline.Factor

min(mask=sentinel('NotSpecified'))#

Create a 1-dimensional factor computing the min of self, each day.

Parameters:: mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when computing results. If supplied, we ignore asset/date pairs where mask produces False.
Returns:: result
Return type:: zipline.pipeline.Factor

median(mask=sentinel('NotSpecified'))#

Create a 1-dimensional factor computing the median of self, each day.

Parameters:: mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when computing results. If supplied, we ignore asset/date pairs where mask produces False.
Returns:: result
Return type:: zipline.pipeline.Factor

sum(mask=sentinel('NotSpecified'))#

Create a 1-dimensional factor computing the sum of self, each day.

Parameters:: mask (zipline.pipeline.Filter, optional) – A Filter representing assets to consider when computing results. If supplied, we ignore asset/date pairs where mask produces False.
Returns:: result
Return type:: zipline.pipeline.Factor

class zipline.pipeline.Term(domain=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), window_safe=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), *args, **kwargs)[source]#

Base class for objects that can appear in the compute graph of a zipline.pipeline.Pipeline.

Notes

Most Pipeline API users only interact with Term via subclasses:

BoundColumn
Factor
Filter
Classifier

Instances of Term are memoized. If you call a Term’s constructor with the same arguments twice, the same object will be returned from both calls:

Example:

>>> from zipline.pipeline.data import EquityPricing
>>> from zipline.pipeline.factors import SimpleMovingAverage
>>> x = SimpleMovingAverage(inputs=[EquityPricing.close], window_length=5)
>>> y = SimpleMovingAverage(inputs=[EquityPricing.close], window_length=5)
>>> x is y
True

Warning

Memoization of terms means that it’s generally unsafe to modify attributes of a term after construction.

graph_repr()[source]#: A short repr to use when rendering GraphViz graphs.

recursive_repr()[source]#: A short repr to use when recursively rendering terms with inputs.

class zipline.pipeline.data.DataSet[source]#

Base class for Pipeline datasets.

A DataSet is defined by two parts:

A collection of Column objects that describe the queryable attributes of the dataset.
A Domain describing the assets and calendar of the data represented by the DataSet.

To create a new Pipeline dataset, define a subclass of DataSet and set one or more Column objects as class-level attributes. Each column requires a np.dtype that describes the type of data that should be produced by a loader for the dataset. Integer columns must also provide a “missing value” to be used when no value is available for a given asset/date combination.

By default, the domain of a dataset is the special singleton value, GENERIC, which means that they can be used in a Pipeline running on any domain.

In some cases, it may be preferable to restrict a dataset to only allow support a single domain. For example, a DataSet may describe data from a vendor that only covers the US. To restrict a dataset to a specific domain, define a domain attribute at class scope.

You can also define a domain-specific version of a generic DataSet by calling its specialize method with the domain of interest.

Examples

The built-in EquityPricing dataset is defined as follows:

class EquityPricing(DataSet):
    open = Column(float)
    high = Column(float)
    low = Column(float)
    close = Column(float)
    volume = Column(float)

The built-in USEquityPricing dataset is a specialization of EquityPricing. It is defined as:

from zipline.pipeline.domain import US_EQUITIES
USEquityPricing = EquityPricing.specialize(US_EQUITIES)

Columns can have types other than float. A dataset containing assorted company metadata might be defined like this:

class CompanyMetadata(DataSet):
    # Use float for semantically-numeric data, even if it's always
    # integral valued (see Notes section below). The default missing
    # value for floats is NaN.
    shares_outstanding = Column(float)

    # Use object for string columns. The default missing value for
    # object-dtype columns is None.
    ticker = Column(object)

    # Use integers for integer-valued categorical data like sector or
    # industry codes. Integer-dtype columns require an explicit missing
    # value.
    sector_code = Column(int, missing_value=-1)

    # Use bool for boolean-valued flags. Note that the default missing
    # value for bool-dtype columns is False.
    is_primary_share = Column(bool)

Notes

Because numpy has no native support for integers with missing values, users are strongly encouraged to use floats for any data that’s semantically numeric. Doing so enables the use of NaN as a natural missing value, which has useful propagation semantics.

classmethod get_column(name)[source]#

Look up a column by name.

Parameters:: name (str) – Name of the column to look up.
Returns:: column – Column with the given name.
Return type:: zipline.pipeline.data.BoundColumn
Raises:: AttributeError – If no column with the given name exists.

class zipline.pipeline.data.Column(dtype, missing_value=sentinel('NotSpecified'), doc=None, metadata=None, currency_aware=False)[source]#

An abstract column of data, not yet associated with a dataset.

bind(name)[source]#: Bind a Column object to its name.

class zipline.pipeline.data.BoundColumn(dtype, missing_value, dataset, name, doc, metadata, currency_conversion, currency_aware)[source]#

A column of data that’s been concretely bound to a particular dataset.

dtype#

The dtype of data produced when this column is loaded.

Type:: numpy.dtype

latest#

A Filter, Factor, or Classifier computing the most recently known value of this column on each date. See zipline.pipeline.mixins.LatestMixin for more details.

Type:: zipline.pipeline.LoadableTerm

dataset#

The dataset to which this column is bound.

Type:: zipline.pipeline.data.DataSet

name#

The name of this column.

Type:: str

metadata#

Extra metadata associated with this column.

Type:: dict

currency_aware#

Whether or not this column produces currency-denominated data.

Type:: bool

Notes

Instances of this class are dynamically created upon access to attributes of DataSet. For example, close is an instance of this class. Pipeline API users should never construct instances of this directly.

property currency_aware#: Whether or not this column produces currency-denominated data.

property currency_conversion#: Specification for currency conversions applied for this term.

property dataset#: The dataset to which this column is bound.

fx(currency)[source]#

Construct a currency-converted version of this column.

Parameters:: currency (str or zipline.currency.Currency) – Currency into which to convert this column’s data.
Returns:: column – Column producing the same data as self, but currency-converted into currency.
Return type:: BoundColumn

graph_repr()[source]#: Short repr to use when rendering Pipeline graphs.

property metadata#: A copy of the metadata for this column.

property name#: The name of this column.

property qualname#: The fully-qualified name of this column.

recursive_repr()[source]#: Short repr used to render in recursive contexts.

specialize(domain)[source]#: Specialize self to a concrete domain.

unspecialize()[source]#

Unspecialize a column to its generic form.

This is equivalent to column.specialize(GENERIC).

class zipline.pipeline.data.DataSetFamily[source]#

Base class for Pipeline dataset families.

Dataset families are used to represent data where the unique identifier for a row requires more than just asset and date coordinates. A DataSetFamily can also be thought of as a collection of DataSet objects, each of which has the same columns, domain, and ndim.

DataSetFamily objects are defined with one or more Column objects, plus one additional field: extra_dims.

The extra_dims field defines coordinates other than asset and date that must be fixed to produce a logical timeseries. The column objects determine columns that will be shared by slices of the family.

extra_dims are represented as an ordered dictionary where the keys are the dimension name, and the values are a set of unique values along that dimension.

To work with a DataSetFamily in a pipeline expression, one must choose a specific value for each of the extra dimensions using the slice() method. For example, given a DataSetFamily:

class SomeDataSet(DataSetFamily):
    extra_dims = [
        ('dimension_0', {'a', 'b', 'c'}),
        ('dimension_1', {'d', 'e', 'f'}),
    ]

    column_0 = Column(float)
    column_1 = Column(bool)

This dataset might represent a table with the following columns:

sid :: int64
asof_date :: datetime64[ns]
timestamp :: datetime64[ns]
dimension_0 :: str
dimension_1 :: str
column_0 :: float64
column_1 :: bool

Here we see the implicit sid, asof_date and timestamp columns as well as the extra dimensions columns.

This DataSetFamily can be converted to a regular DataSet with:

DataSetSlice = SomeDataSet.slice(dimension_0='a', dimension_1='e')

This sliced dataset represents the rows from the higher dimensional dataset where (dimension_0 == 'a') & (dimension_1 == 'e').

classmethod slice(*args, **kwargs)[source]#

Take a slice of a DataSetFamily to produce a dataset indexed by asset and date.

Parameters:

*args –
**kwargs – The coordinates to fix along each extra dimension.

Returns:

dataset – A regular pipeline dataset indexed by asset and date.

Return type:

DataSet

Notes

The extra dimensions coords used to produce the result are available under the extra_coords attribute.

class zipline.pipeline.data.EquityPricing[source]#

DataSet containing daily trading prices and volumes.

close = EquityPricing.close::float64#

high = EquityPricing.high::float64#

low = EquityPricing.low::float64#

open = EquityPricing.open::float64#

volume = EquityPricing.volume::float64#

Built-in Factors#

Factors aim to transform the input data in a way that extracts a signal on which the algorithm can trade.

class zipline.pipeline.factors.AverageDollarVolume(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Average Daily Dollar Volume

Default Inputs: [EquityPricing.close, EquityPricing.volume]

Default Window Length: None

compute(today, assets, out, close, volume)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.BollingerBands(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Bollinger Bands technical indicator. https://en.wikipedia.org/wiki/Bollinger_Bands

Default Inputs: zipline.pipeline.data.EquityPricing.close

Parameters:

inputs (length-1 iterable[BoundColumn]) – The expression over which to compute bollinger bands.
window_length (int > 0) – Length of the lookback window over which to compute the bollinger bands.
k (float) – The number of standard deviations to add or subtract to create the upper and lower bands.

compute(today, assets, out, close, k)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.BusinessDaysSincePreviousEvent(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), domain=sentinel('NotSpecified'), *args, **kwargs)[source]#

Abstract class for business days since a previous event. Returns the number of business days (not trading days!) since the most recent event date for each asset.

This doesn’t use trading days for symmetry with BusinessDaysUntilNextEarnings.

Assets which announced or will announce the event today will produce a value of 0.0. Assets that announced the event on the previous business day will produce a value of 1.0.

Assets for which the event date is NaT will produce a value of NaN.

Example

BusinessDaysSincePreviousEvent can be used to create an event-driven factor. For instance, you may want to only trade assets that have a data point with an asof_date in the last 5 business days. To do this, you can create a BusinessDaysSincePreviousEvent factor, supplying the relevant asof_date column from your dataset as input, like this:

# Factor computing number of days since most recent asof_date
# per asset.
days_since_event = BusinessDaysSincePreviousEvent(
    inputs=[MyDataset.asof_date]
)

# Filter returning True for each asset whose most recent asof_date
# was in the last 5 business days.
recency_filter = (days_since_event <= 5)

dtype = dtype('float64')#

class zipline.pipeline.factors.BusinessDaysUntilNextEvent(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), domain=sentinel('NotSpecified'), *args, **kwargs)[source]#

Abstract class for business days since a next event. Returns the number of business days (not trading days!) until the next known event date for each asset.

This doesn’t use trading days because the trading calendar includes information that may not have been available to the algorithm at the time when compute is called.

For example, the NYSE closings September 11th 2001, would not have been known to the algorithm on September 10th.

Assets that announced or will announce the event today will produce a value of 0.0. Assets that will announce the event on the next upcoming business day will produce a value of 1.0.

Assets for which the event date is NaT will produce a value of NaN.

dtype = dtype('float64')#

class zipline.pipeline.factors.DailyReturns(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Calculates daily percent change in close price.

Default Inputs: [EquityPricing.close]

class zipline.pipeline.factors.ExponentialWeightedMovingAverage(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Exponentially Weighted Moving Average

Default Inputs: None

Default Window Length: None

Parameters:

inputs (length-1 list/tuple of BoundColumn) – The expression over which to compute the average.
window_length (int > 0) – Length of the lookback window over which to compute the average.
decay_rate (float, 0 < decay_rate <= 1) –
Weighting factor by which to discount past observations.

When calculating historical averages, rows are multiplied by the sequence:
```
decay_rate, decay_rate ** 2, decay_rate ** 3, ...
```

Notes

This class can also be imported under the name EWMA.

See also

pandas.DataFrame.ewm()

compute(today, assets, out, data, decay_rate)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.ExponentialWeightedMovingStdDev(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Exponentially Weighted Moving Standard Deviation

Default Inputs: None

Default Window Length: None

Parameters:

inputs (length-1 list/tuple of BoundColumn) – The expression over which to compute the average.
window_length (int > 0) – Length of the lookback window over which to compute the average.
decay_rate (float, 0 < decay_rate <= 1) –
Weighting factor by which to discount past observations.

When calculating historical averages, rows are multiplied by the sequence:
```
decay_rate, decay_rate ** 2, decay_rate ** 3, ...
```

Notes

This class can also be imported under the name EWMSTD.

See also

pandas.DataFrame.ewm()

compute(today, assets, out, data, decay_rate)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.Latest(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Factor producing the most recently-known value of inputs[0] on each day.

The .latest attribute of DataSet columns returns an instance of this Factor.

compute(today, assets, out, data)[source]#: Override this method with a function that writes a value into out.

zipline.pipeline.factors.MACDSignal#: alias of MovingAverageConvergenceDivergenceSignal

class zipline.pipeline.factors.MaxDrawdown(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Max Drawdown

Default Inputs: None

Default Window Length: None

compute(today, assets, out, data)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.Returns(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Calculates the percent change in close price over the given window_length.

Default Inputs: [EquityPricing.close]

compute(today, assets, out, close)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.RollingPearson(base_factor, target, correlation_length, mask=sentinel('NotSpecified'))[source]#

A Factor that computes pearson correlation coefficients between the columns of a given Factor and either the columns of another Factor/BoundColumn or a slice/single column of data.

Parameters:

base_factor (zipline.pipeline.Factor) – The factor for which to compute correlations of each of its columns with target.
target (zipline.pipeline.Term with a numeric dtype) – The term with which to compute correlations against each column of data produced by base_factor. This term may be a Factor, a BoundColumn or a Slice. If target is two-dimensional, correlations are computed asset-wise.
correlation_length (int) – Length of the lookback window over which to compute each correlation coefficient.
mask (zipline.pipeline.Filter, optional) – A Filter describing which assets (columns) of base_factor should have their correlation with target computed each day.

Notes

Most users should call Factor.pearsonr rather than directly construct an instance of this class.

compute(today, assets, out, base_data, target_data)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.RollingSpearman(base_factor, target, correlation_length, mask=sentinel('NotSpecified'))[source]#

A Factor that computes spearman rank correlation coefficients between the columns of a given Factor and either the columns of another Factor/BoundColumn or a slice/single column of data.

Parameters:

base_factor (zipline.pipeline.Factor) – The factor for which to compute correlations of each of its columns with target.
target (zipline.pipeline.Term with a numeric dtype) – The term with which to compute correlations against each column of data produced by base_factor. This term may be a Factor, a BoundColumn or a Slice. If target is two-dimensional, correlations are computed asset-wise.
correlation_length (int) – Length of the lookback window over which to compute each correlation coefficient.
mask (zipline.pipeline.Filter, optional) – A Filter describing which assets (columns) of base_factor should have their correlation with target computed each day.

Notes

Most users should call Factor.spearmanr rather than directly construct an instance of this class.

compute(today, assets, out, base_data, target_data)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.RollingLinearRegressionOfReturns(target, returns_length, regression_length, mask=sentinel('NotSpecified'))[source]#

Perform an ordinary least-squares regression predicting the returns of all other assets on the given asset.

Parameters:

target (zipline.assets.Asset) – The asset to regress against all other assets.
returns_length (int >= 2) – Length of the lookback window over which to compute returns. Daily returns require a window length of 2.
regression_length (int >= 1) – Length of the lookback window over which to compute each regression.
mask (zipline.pipeline.Filter, optional) – A Filter describing which assets should be regressed against the target asset each day.

Notes

Computing this factor over many assets can be time consuming. It is recommended that a mask be used in order to limit the number of assets over which regressions are computed.

This factor is designed to return five outputs:

alpha, a factor that computes the intercepts of each regression.
beta, a factor that computes the slopes of each regression.
r_value, a factor that computes the correlation coefficient of each regression.
p_value, a factor that computes, for each regression, the two-sided p-value for a hypothesis test whose null hypothesis is that the slope is zero.
stderr, a factor that computes the standard error of the estimate of each regression.

For more help on factors with multiple outputs, see zipline.pipeline.CustomFactor.

Examples

Let the following be example 10-day returns for three different assets:

               SPY    MSFT     FB
2017-03-13    -.03     .03    .04
2017-03-14    -.02    -.03    .02
2017-03-15    -.01     .02    .01
2017-03-16       0    -.02    .01
2017-03-17     .01     .04   -.01
2017-03-20     .02    -.03   -.02
2017-03-21     .03     .01   -.02
2017-03-22     .04    -.02   -.02

Suppose we are interested in predicting each stock’s returns from SPY’s over rolling 5-day look back windows. We can compute rolling regression coefficients (alpha and beta) from 2017-03-17 to 2017-03-22 by doing:

regression_factor = RollingRegressionOfReturns(
    target=sid(8554),
    returns_length=10,
    regression_length=5,
)
alpha = regression_factor.alpha
beta = regression_factor.beta

The result of computing alpha from 2017-03-17 to 2017-03-22 gives:

               SPY    MSFT     FB
2017-03-17       0    .011   .003
2017-03-20       0   -.004   .004
2017-03-21       0    .007   .006
2017-03-22       0    .002   .008

And the result of computing beta from 2017-03-17 to 2017-03-22 gives:

               SPY    MSFT     FB
2017-03-17       1      .3   -1.1
2017-03-20       1      .2     -1
2017-03-21       1     -.3     -1
2017-03-22       1     -.3    -.9

Note that SPY’s column for alpha is all 0’s and for beta is all 1’s, as the regression line of SPY with itself is simply the function y = x.

To understand how each of the other values were calculated, take for example MSFT’s alpha and beta values on 2017-03-17 (.011 and .3, respectively). These values are the result of running a linear regression predicting MSFT’s returns from SPY’s returns, using values starting at 2017-03-17 and looking back 5 days. That is, the regression was run with x = [-.03, -.02, -.01, 0, .01] and y = [.03, -.03, .02, -.02, .04], and it produced a slope of .3 and an intercept of .011.

class zipline.pipeline.factors.RollingPearsonOfReturns(target, returns_length, correlation_length, mask=sentinel('NotSpecified'))[source]#

Calculates the Pearson product-moment correlation coefficient of the returns of the given asset with the returns of all other assets.

Pearson correlation is what most people mean when they say “correlation coefficient” or “R-value”.

Parameters:

target (zipline.assets.Asset) – The asset to correlate with all other assets.
returns_length (int >= 2) – Length of the lookback window over which to compute returns. Daily returns require a window length of 2.
correlation_length (int >= 1) – Length of the lookback window over which to compute each correlation coefficient.
mask (zipline.pipeline.Filter, optional) – A Filter describing which assets should have their correlation with the target asset computed each day.

Notes

Computing this factor over many assets can be time consuming. It is recommended that a mask be used in order to limit the number of assets over which correlations are computed.

Examples

Let the following be example 10-day returns for three different assets:

               SPY    MSFT     FB
2017-03-13    -.03     .03    .04
2017-03-14    -.02    -.03    .02
2017-03-15    -.01     .02    .01
2017-03-16       0    -.02    .01
2017-03-17     .01     .04   -.01
2017-03-20     .02    -.03   -.02
2017-03-21     .03     .01   -.02
2017-03-22     .04    -.02   -.02

Suppose we are interested in SPY’s rolling returns correlation with each stock from 2017-03-17 to 2017-03-22, using a 5-day look back window (that is, we calculate each correlation coefficient over 5 days of data). We can achieve this by doing:

rolling_correlations = RollingPearsonOfReturns(
    target=sid(8554),
    returns_length=10,
    correlation_length=5,
)

The result of computing rolling_correlations from 2017-03-17 to 2017-03-22 gives:

               SPY   MSFT     FB
2017-03-17       1    .15   -.96
2017-03-20       1    .10   -.96
2017-03-21       1   -.16   -.94
2017-03-22       1   -.16   -.85

Note that the column for SPY is all 1’s, as the correlation of any data series with itself is always 1. To understand how each of the other values were calculated, take for example the .15 in MSFT’s column. This is the correlation coefficient between SPY’s returns looking back from 2017-03-17 (-.03, -.02, -.01, 0, .01) and MSFT’s returns (.03, -.03, .02, -.02, .04).

class zipline.pipeline.factors.RollingSpearmanOfReturns(target, returns_length, correlation_length, mask=sentinel('NotSpecified'))[source]#

Calculates the Spearman rank correlation coefficient of the returns of the given asset with the returns of all other assets.

Parameters:

target (zipline.assets.Asset) – The asset to correlate with all other assets.
returns_length (int >= 2) – Length of the lookback window over which to compute returns. Daily returns require a window length of 2.
correlation_length (int >= 1) – Length of the lookback window over which to compute each correlation coefficient.
mask (zipline.pipeline.Filter, optional) – A Filter describing which assets should have their correlation with the target asset computed each day.

Notes

Computing this factor over many assets can be time consuming. It is recommended that a mask be used in order to limit the number of assets over which correlations are computed.

class zipline.pipeline.factors.SimpleBeta(target, regression_length, allowed_missing_percentage=0.25)[source]#

Factor producing the slope of a regression line between each asset’s daily returns to the daily returns of a single “target” asset.

Parameters:

target (zipline.Asset) – Asset against which other assets should be regressed.
regression_length (int) – Number of days of daily returns to use for the regression.
allowed_missing_percentage (float, optional) – Percentage of returns observations (between 0 and 1) that are allowed to be missing when calculating betas. Assets with more than this percentage of returns observations missing will produce values of NaN. Default behavior is that 25% of inputs can be missing.

compute(today, assets, out, all_returns, target_returns, allowed_missing_count)[source]#: Override this method with a function that writes a value into out.

dtype = dtype('float64')#

graph_repr()[source]#: Short repr to use when rendering Pipeline graphs.

property target#: Get the target of the beta calculation.

class zipline.pipeline.factors.RSI(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Relative Strength Index

Default Inputs: zipline.pipeline.data.EquityPricing.close

Default Window Length: 15

compute(today, assets, out, closes)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.SimpleMovingAverage(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Average Value of an arbitrary column

Default Inputs: None

Default Window Length: None

compute(today, assets, out, data)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.VWAP(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Volume Weighted Average Price

Default Inputs: [EquityPricing.close, EquityPricing.volume]

Default Window Length: None

class zipline.pipeline.factors.WeightedAverageValue(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Helper for VWAP-like computations.

Default Inputs: None

Default Window Length: None

compute(today, assets, out, base, weight)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.PercentChange(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Calculates the percent change over the given window_length.

Default Inputs: None

Default Window Length: None

Notes

Percent change is calculated as (new - old) / abs(old).

compute(today, assets, out, values)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.factors.PeerCount(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Peer Count of distinct categories in a given classifier. This factor is returned by the classifier instance method peer_count()

Default Inputs: None

Default Window Length: 1

compute(today, assets, out, classifier_values)[source]#: Override this method with a function that writes a value into out.

Built-in Filters#

class zipline.pipeline.filters.All(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

A Filter requiring that assets produce True for window_length consecutive days.

Default Inputs: None

Default Window Length: None

compute(today, assets, out, arg)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.filters.AllPresent(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

Pipeline filter indicating input term has data for a given window.

compute(today, assets, out, value)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.filters.Any(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

A Filter requiring that assets produce True for at least one day in the last window_length days.

Default Inputs: None

Default Window Length: None

compute(today, assets, out, arg)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.filters.AtLeastN(inputs=sentinel('NotSpecified'), outputs=sentinel('NotSpecified'), window_length=sentinel('NotSpecified'), mask=sentinel('NotSpecified'), dtype=sentinel('NotSpecified'), missing_value=sentinel('NotSpecified'), ndim=sentinel('NotSpecified'), **kwargs)[source]#

A Filter requiring that assets produce True for at least N days in the last window_length days.

Default Inputs: None

Default Window Length: None

compute(today, assets, out, arg, N)[source]#: Override this method with a function that writes a value into out.

class zipline.pipeline.filters.SingleAsset(asset)[source]#

A Filter that computes to True only for the given asset.

graph_repr()[source]#: A short repr to use when rendering GraphViz graphs.

class zipline.pipeline.filters.StaticAssets(assets)[source]#

A Filter that computes True for a specific set of predetermined assets.

StaticAssets is mostly useful for debugging or for interactively computing pipeline terms for a fixed set of assets that are known ahead of time.

Parameters:: assets (iterable[Asset]) – An iterable of assets for which to filter.

class zipline.pipeline.filters.StaticSids(sids)[source]#

A Filter that computes True for a specific set of predetermined sids.

StaticSids is mostly useful for debugging or for interactively computing pipeline terms for a fixed set of sids that are known ahead of time.

Parameters:: sids (iterable[int]) – An iterable of sids for which to filter.

Pipeline Engine#

Computation engines for executing a Pipeline define the core computation algorithms.

The primary entrypoint is SimplePipelineEngine.run_pipeline, which implements the following algorithm for executing pipelines:

Determine the domain of the pipeline.
Build a dependency graph of all terms in pipeline, with information about how many extra rows each term needs from its inputs.
Combine the domain computed in (2) with our AssetFinder to produce a “lifetimes matrix”. The lifetimes matrix is a DataFrame of booleans whose labels are dates x assets. Each entry corresponds to a (date, asset) pair and indicates whether the asset in question was tradable on the date in question.
Produce a “workspace” dictionary with cached or otherwise pre-computed terms.
Topologically sort the graph constructed in (1) to produce an execution order for any terms that were not pre-populated.
Iterate over the terms in the order computed in (5). For each term:
1. Fetch the term’s inputs from the workspace.
2. Compute each term and store the results in the workspace.
3. Remove the results from the workspace if their are no longer needed to reduce memory use during execution.
Extract the pipeline’s outputs from the workspace and convert them into “narrow” format, with output labels dictated by the Pipeline’s screen.

class zipline.pipeline.engine.PipelineEngine[source]#

abstract run_pipeline(pipeline, start_date, end_date, hooks=None)[source]#

Compute values for pipeline from start_date to end_date.

Parameters:

pipeline (zipline.pipeline.Pipeline) – The pipeline to run.
start_date (pd.Timestamp) – Start date of the computed matrix.
end_date (pd.Timestamp) – End date of the computed matrix.
hooks (list[implements(PipelineHooks)], optional) – Hooks for instrumenting Pipeline execution.

Returns:

result – A frame of computed results.

The result columns correspond to the entries of pipeline.columns, which should be a dictionary mapping strings to instances of zipline.pipeline.Term.

For each date between start_date and end_date, result will contain a row for each asset that passed pipeline.screen. A screen of None indicates that a row should be returned for each asset that existed each day.

Return type:

pd.DataFrame

abstract run_chunked_pipeline(pipeline, start_date, end_date, chunksize, hooks=None)[source]#

Compute values for pipeline from start_date to end_date, in date chunks of size chunksize.

Chunked execution reduces memory consumption, and may reduce computation time depending on the contents of your pipeline.

Parameters:

pipeline (Pipeline) – The pipeline to run.
start_date (pd.Timestamp) – The start date to run the pipeline for.
end_date (pd.Timestamp) – The end date to run the pipeline for.
chunksize (int) – The number of days to execute at a time.
hooks (list[implements(PipelineHooks)], optional) – Hooks for instrumenting Pipeline execution.

Returns:

result – A frame of computed results.

The result columns correspond to the entries of pipeline.columns, which should be a dictionary mapping strings to instances of zipline.pipeline.Term.

Return type:

pd.DataFrame

class zipline.pipeline.engine.SimplePipelineEngine(get_loader, asset_finder, default_domain=GENERIC, populate_initial_workspace=None, default_hooks=None)[source]#

PipelineEngine class that computes each term independently.

Parameters:

get_loader (callable) – A function that is given a loadable term and returns a PipelineLoader to use to retrieve raw data for that term.
asset_finder (zipline.assets.AssetFinder) – An AssetFinder instance. We depend on the AssetFinder to determine which assets are in the top-level universe at any point in time.
populate_initial_workspace (callable, optional) – A function which will be used to populate the initial workspace when computing a pipeline. See zipline.pipeline.engine.default_populate_initial_workspace() for more info.
default_hooks (list, optional) – List of hooks that should be used to instrument all pipelines executed by this engine.

__init__(get_loader, asset_finder, default_domain=GENERIC, populate_initial_workspace=None, default_hooks=None)[source]#

run_chunked_pipeline(pipeline, start_date, end_date, chunksize, hooks=None)[source]#

Compute values for pipeline from start_date to end_date, in date chunks of size chunksize.

Chunked execution reduces memory consumption, and may reduce computation time depending on the contents of your pipeline.

Parameters:

pipeline (Pipeline) – The pipeline to run.
start_date (pd.Timestamp) – The start date to run the pipeline for.
end_date (pd.Timestamp) – The end date to run the pipeline for.
chunksize (int) – The number of days to execute at a time.
hooks (list[implements(PipelineHooks)], optional) – Hooks for instrumenting Pipeline execution.

Returns:

result – A frame of computed results.

The result columns correspond to the entries of pipeline.columns, which should be a dictionary mapping strings to instances of zipline.pipeline.Term.

Return type:

pd.DataFrame

run_pipeline(pipeline, start_date, end_date, hooks=None)[source]#

Compute values for pipeline from start_date to end_date.

Parameters:

pipeline (zipline.pipeline.Pipeline) – The pipeline to run.
start_date (pd.Timestamp) – Start date of the computed matrix.
end_date (pd.Timestamp) – End date of the computed matrix.
hooks (list[implements(PipelineHooks)], optional) – Hooks for instrumenting Pipeline execution.

Returns:

result – A frame of computed results.

The result columns correspond to the entries of pipeline.columns, which should be a dictionary mapping strings to instances of zipline.pipeline.Term.

Return type:

pd.DataFrame

zipline.pipeline.engine.default_populate_initial_workspace(initial_workspace, root_mask_term, execution_plan, dates, assets)[source]#

The default implementation for populate_initial_workspace. This function returns the initial_workspace argument without making any modifications.

Parameters:

initial_workspace (dict[array-like]) – The initial workspace before we have populated it with any cached terms.
root_mask_term (Term) – The root mask term, normally AssetExists(). This is needed to compute the dates for individual terms.
execution_plan (ExecutionPlan) – The execution plan for the pipeline being run.
dates (pd.DatetimeIndex) – All of the dates being requested in this pipeline run including the extra dates for look back windows.
assets (pd.Int64Index) – All of the assets that exist for the window being computed.

Returns:

populated_initial_workspace – The workspace to begin computations with.

Return type:

dict[term, array-like]

Data Loaders#

There are several loaders to feed data to a Pipeline that need to implement the interface defined by the PipelineLoader.

class zipline.pipeline.loaders.base.PipelineLoader(*args, **kwargs)[source]#

Interface for PipelineLoaders.

load_adjusted_array(domain, columns, dates, sids, mask)[source]#

Load data for columns as AdjustedArrays.

Parameters:

domain (zipline.pipeline.domain.Domain) – The domain of the pipeline for which the requested data must be loaded.
columns (list[zipline.pipeline.data.dataset.BoundColumn]) – Columns for which data is being requested.
dates (pd.DatetimeIndex) – Dates for which data is being requested.
sids (pd.Int64Index) – Asset identifiers for which data is being requested.
mask (np.array[ndim=2, dtype=bool]) – Boolean array of shape (len(dates), len(sids)) indicating dates on which we believe the requested assets were alive/tradeable. This is used for optimization by some loaders.

Returns:

arrays – Map from column to an AdjustedArray representing a point-in-time rolling view over the requested dates for the requested sids.

Return type:

dict[BoundColumn -> zipline.lib.adjusted_array.AdjustedArray]

__init__()#

class zipline.pipeline.loaders.frame.DataFrameLoader(column, baseline, adjustments=None)[source]#

A PipelineLoader that reads its input from DataFrames.

Mostly useful for testing, but can also be used for real work if your data fits in memory.

Parameters:

column (zipline.pipeline.data.BoundColumn) – The column whose data is loadable by this loader.
baseline (pandas.DataFrame) – A DataFrame with index of type DatetimeIndex and columns of type Int64Index. Dates should be labelled with the first date on which a value would be available to an algorithm. This means that OHLCV data should generally be shifted back by a trading day before being supplied to this class.
adjustments (pandas.DataFrame, default=None) –

A DataFrame with the following columns:
sid : int value : any kind : int (zipline.pipeline.loaders.frame.ADJUSTMENT_TYPES) start_date : datetime64 (can be NaT) end_date : datetime64 (must be set) apply_date : datetime64 (must be set)

The default of None is interpreted as “no adjustments to the baseline”.

__init__(column, baseline, adjustments=None)[source]#

format_adjustments(dates, assets)[source]#

Build a dict of Adjustment objects in the format expected by AdjustedArray.

Returns a dict of the form: { # Integer index into dates for the date on which we should # apply the list of adjustments. 1 : [ Float64Multiply(first_row=2, last_row=4, col=3, value=0.5), Float64Overwrite(first_row=3, last_row=5, col=1, value=2.0), … ], … }

load_adjusted_array(domain, columns, dates, sids, mask)[source]#: Load data from our stored baseline.

class zipline.pipeline.loaders.equity_pricing_loader.EquityPricingLoader(raw_price_reader, adjustments_reader, fx_reader)[source]#

A PipelineLoader for loading daily OHLCV data.

Parameters:

raw_price_reader (zipline.data.session_bars.SessionBarReader) – Reader providing raw prices.
adjustments_reader (zipline.data.adjustments.SQLiteAdjustmentReader) – Reader providing price/volume adjustments.
fx_reader (zipline.data.fx.FXRateReader) – Reader providing currency conversions.

__init__(raw_price_reader, adjustments_reader, fx_reader)[source]#

zipline.pipeline.loaders.equity_pricing_loader.USEquityPricingLoader#: alias of EquityPricingLoader

class zipline.pipeline.loaders.events.EventsLoader(events, next_value_columns, previous_value_columns)[source]#

Base class for PipelineLoaders that supports loading the next and previous value of an event field.

Does not currently support adjustments.

Parameters:

events (pd.DataFrame) –
A DataFrame representing events (e.g. share buybacks or earnings announcements) associated with particular companies.

events must contain at least three columns::

sidint64
The asset id associated with each event.

event_datedatetime64[ns]
The date on which the event occurred.

timestampdatetime64[ns]
The date on which we learned about the event.
next_value_columns (dict[BoundColumn -> str]) – Map from dataset columns to raw field names that should be used when searching for a next event value.
previous_value_columns (dict[BoundColumn -> str]) – Map from dataset columns to raw field names that should be used when searching for a previous event value.

__init__(events, next_value_columns, previous_value_columns)[source]#

class zipline.pipeline.loaders.earnings_estimates.EarningsEstimatesLoader(estimates, name_map)[source]#

An abstract pipeline loader for estimates data that can load data a variable number of quarters forwards/backwards from calendar dates depending on the num_announcements attribute of the columns’ dataset. If split adjustments are to be applied, a loader, split-adjusted columns, and the split-adjusted asof-date must be supplied.

Parameters:

estimates (pd.DataFrame) –

The raw estimates data; must contain at least 5 columns:

sidint64
The asset id associated with each estimate.

event_datedatetime64[ns]
The date on which the event that the estimate is for will/has occurred.

timestampdatetime64[ns]
The datetime where we learned about the estimate.

fiscal_quarterint64
The quarter during which the event has/will occur.

fiscal_yearint64
The year during which the event has/will occur.
name_map (dict[str -> str]) – A map of names of BoundColumns that this loader will load to the names of the corresponding columns in events.

__init__(estimates, name_map)[source]#

Exchange and Asset Metadata#

class zipline.assets.ExchangeInfo(name, canonical_name, country_code)[source]#

An exchange where assets are traded.

Parameters:

name (str or None) – The full name of the exchange, for example ‘NEW YORK STOCK EXCHANGE’ or ‘NASDAQ GLOBAL MARKET’.
canonical_name (str) – The canonical name of the exchange, for example ‘NYSE’ or ‘NASDAQ’. If None this will be the same as the name.
country_code (str) – The country code where the exchange is located.

name#

The full name of the exchange, for example ‘NEW YORK STOCK EXCHANGE’ or ‘NASDAQ GLOBAL MARKET’.

Type:: str or None

canonical_name#

The canonical name of the exchange, for example ‘NYSE’ or ‘NASDAQ’. If None this will be the same as the name.

Type:: str

country_code#

The country code where the exchange is located.

Type:: str

calendar#

The trading calendar the exchange uses.

Type:: TradingCalendar

property calendar#: The trading calendar that this exchange uses.

class zipline.assets.Asset#

Base class for entities that can be owned by a trading algorithm.

sid#

Persistent unique identifier assigned to the asset.

Type:: int

symbol#

Most recent ticker under which the asset traded. This field can change without warning if the asset changes tickers. Use sid if you need a persistent identifier.

Type:: str

asset_name#

Full name of the asset.

Type:: str

exchange#

Canonical short name of the exchange on which the asset trades (e.g., ‘NYSE’).

Type:: str

exchange_full#

Full name of the exchange on which the asset trades (e.g., ‘NEW YORK STOCK EXCHANGE’).

Type:: str

exchange_info#

Information about the exchange this asset is listed on.

Type:: zipline.assets.ExchangeInfo

country_code#

Two character code indicating the country in which the asset trades.

Type:: str

start_date#

Date on which the asset first traded.

Type:: pd.Timestamp

end_date#

Last date on which the asset traded. On Quantopian, this value is set to the current (real time) date for assets that are still trading.

Type:: pd.Timestamp

tick_size#

Minimum amount that the price can change for this asset.

Type:: float

auto_close_date#

Date on which positions in this asset will be automatically liquidated to cash during a simulation. By default, this is three days after end_date.

Type:: pd.Timestamp

from_dict()#: Build an Asset instance from a dict.

is_alive_for_session()#

Returns whether the asset is alive at the given dt.

Parameters:: session_label (pd.Timestamp) – The desired session label to check. (midnight UTC)
Returns:: boolean
Return type:: whether the asset is alive at the given dt.

is_exchange_open()#

Parameters:: dt_minute (pd.Timestamp (UTC, tz-aware)) – The minute to check.
Returns:: boolean
Return type:: whether the asset’s exchange is open at the given minute.

to_dict()#

Convert to a python dict containing all attributes of the asset.

This is often useful for debugging.

Returns:: as_dict
Return type:: dict

class zipline.assets.Equity#: Asset subclass representing partial ownership of a company, trust, or partnership.

class zipline.assets.Future#

Asset subclass representing ownership of a futures contract.

to_dict()#: Convert to a python dict.

class zipline.assets.AssetConvertible[source]#

ABC for types that are convertible to integer-representations of Assets.

Includes Asset, str, and Integral

Trading Calendar API#

The events that generate the timeline of the algorithm execution adhere to a given TradingCalendar.

Data API#

Writers#

class zipline.data.bcolz_daily_bars.BcolzDailyBarWriter(filename, calendar, start_session, end_session)[source]#

Class capable of writing daily OHLCV data to disk in a format that can be read efficiently by BcolzDailyOHLCVReader.

Parameters:

filename (str) – The location at which we should write our output.
calendar (zipline.utils.calendar.trading_calendar) – Calendar to use to compute asset calendar offsets.
start_session (pd.Timestamp) – Midnight UTC session label.
end_session (pd.Timestamp) – Midnight UTC session label.

write(data, assets=None, show_progress=False, invalid_data_behavior='warn')[source]#

Parameters:

data (iterable[tuple[int, pandas.DataFrame or bcolz.ctable]]) – The data chunks to write. Each chunk should be a tuple of sid and the data for that asset.
assets (set[int], optional) – The assets that should be in data. If this is provided we will check data against the assets and provide better progress information.
show_progress (bool, optional) – Whether or not to show a progress bar while writing.
invalid_data_behavior ({'warn', 'raise', 'ignore'}, optional) – What to do when data is encountered that is outside the range of a uint32.

Returns:

table – The newly-written table.

Return type:

bcolz.ctable

write_csvs(asset_map, show_progress=False, invalid_data_behavior='warn')[source]#

Read CSVs as DataFrames from our asset map.

Parameters:

asset_map (dict[int -> str]) – A mapping from asset id to file path with the CSV data for that asset
show_progress (bool) – Whether or not to show a progress bar while writing.
invalid_data_behavior ({'warn', 'raise', 'ignore'}) – What to do when data is encountered that is outside the range of a uint32.

class zipline.data.adjustments.SQLiteAdjustmentWriter(conn_or_path, equity_daily_bar_reader, overwrite=False)[source]#

Writer for data to be read by SQLiteAdjustmentReader

Parameters:

conn_or_path (str or sqlite3.Connection) – A handle to the target sqlite database.
equity_daily_bar_reader (SessionBarReader) – Daily bar reader to use for dividend writes.
overwrite (bool, optional, default=False) – If True and conn_or_path is a string, remove any existing files at the given path before connecting.

calc_dividend_ratios(dividends)[source]#

Calculate the ratios to apply to equities when looking back at pricing history so that the price is smoothed over the ex_date, when the market adjusts to the change in equity value due to upcoming dividend.

Returns:: A frame in the same format as splits and mergers, with keys - sid, the id of the equity - effective_date, the date in seconds on which to apply the ratio. - ratio, the ratio to apply to backwards looking pricing data.
Return type:: DataFrame

write(splits=None, mergers=None, dividends=None, stock_dividends=None)[source]#

Writes data to a SQLite file to be read by SQLiteAdjustmentReader.

Parameters:

splits (pandas.DataFrame, optional) –

Dataframe containing split data. The format of this dataframe is:

effective_dateint
The date, represented as seconds since Unix epoch, on which the adjustment should be applied.

ratiofloat
A value to apply to all data earlier than the effective date. For open, high, low, and close those values are multiplied by the ratio. Volume is divided by this value.

sidint
The asset id associated with this adjustment.
mergers (pandas.DataFrame, optional) –

DataFrame containing merger data. The format of this dataframe is:

effective_dateint
The date, represented as seconds since Unix epoch, on which the adjustment should be applied.

ratiofloat
A value to apply to all data earlier than the effective date. For open, high, low, and close those values are multiplied by the ratio. Volume is unaffected.

sidint
The asset id associated with this adjustment.
dividends (pandas.DataFrame, optional) –

DataFrame containing dividend data. The format of the dataframe is:

sidint
The asset id associated with this adjustment.

ex_datedatetime64
The date on which an equity must be held to be eligible to receive payment.

declared_datedatetime64
The date on which the dividend is announced to the public.

pay_datedatetime64
The date on which the dividend is distributed.

record_datedatetime64
The date on which the stock ownership is checked to determine distribution of dividends.

amountfloat
The cash amount paid for each share.

Dividend ratios are calculated as: 1.0 - (dividend_value / "close on day prior to ex_date")
stock_dividends (pandas.DataFrame, optional) –
DataFrame containing stock dividend data. The format of the dataframe is:

sidint
The asset id associated with this adjustment.

ex_datedatetime64
The date on which an equity must be held to be eligible to receive payment.

declared_datedatetime64
The date on which the dividend is announced to the public.

pay_datedatetime64
The date on which the dividend is distributed.

record_datedatetime64
The date on which the stock ownership is checked to determine distribution of dividends.

payment_sidint
The asset id of the shares that should be paid instead of cash.

ratiofloat
The ratio of currently held shares in the held sid that should be paid with new shares of the payment_sid.

write_dividend_data(dividends, stock_dividends=None)[source]#: Write both dividend payouts and the derived price adjustment ratios.

write_dividend_payouts(frame)[source]#: Write dividend payout data to SQLite table dividend_payouts.

class zipline.assets.AssetDBWriter(engine)[source]#

Class used to write data to an assets db.

Parameters:: engine (Engine or str) – An SQLAlchemy engine or path to a SQL database.

init_db(txn=None)[source]#

Connect to database and create tables.

Parameters:: txn (sa.engine.Connection, optional) – The transaction block to execute in. If this is not provided, a new transaction will be started with the engine provided.
Returns:: metadata – The metadata that describes the new assets db.
Return type:: sa.MetaData

write(equities=None, futures=None, exchanges=None, root_symbols=None, equity_supplementary_mappings=None, chunk_size=999)[source]#

Write asset metadata to a sqlite database.

Parameters:

equities (pd.DataFrame, optional) –
The equity metadata. The columns for this dataframe are:

symbolstr
The ticker symbol for this equity.

asset_namestr
The full name for this asset.

start_datedatetime
The date when this asset was created.

end_datedatetime, optional
The last date we have trade data for this asset.

first_tradeddatetime, optional
The first date we have trade data for this asset.

auto_close_datedatetime, optional
The date on which to close any positions in this asset.

exchangestr
The exchange where this asset is traded.

The index of this dataframe should contain the sids.
futures (pd.DataFrame, optional) –
The future contract metadata. The columns for this dataframe are:

symbolstr
The ticker symbol for this futures contract.

root_symbolstr
The root symbol, or the symbol with the expiration stripped out.

asset_namestr
The full name for this asset.

start_datedatetime, optional
The date when this asset was created.

end_datedatetime, optional
The last date we have trade data for this asset.

first_tradeddatetime, optional
The first date we have trade data for this asset.

exchangestr
The exchange where this asset is traded.

notice_datedatetime
The date when the owner of the contract may be forced to take physical delivery of the contract’s asset.

expiration_datedatetime
The date when the contract expires.

auto_close_datedatetime
The date when the broker will automatically close any positions in this contract.

tick_sizefloat
The minimum price movement of the contract.

multiplier: float
The amount of the underlying asset represented by this contract.
exchanges (pd.DataFrame, optional) –
The exchanges where assets can be traded. The columns of this dataframe are:

exchangestr
The full name of the exchange.

canonical_namestr
The canonical name of the exchange.

country_codestr
The ISO 3166 alpha-2 country code of the exchange.
root_symbols (pd.DataFrame, optional) –
The root symbols for the futures contracts. The columns for this dataframe are:

root_symbolstr
The root symbol name.

root_symbol_idint
The unique id for this root symbol.

sectorstring, optional
The sector of this root symbol.

descriptionstring, optional
A short description of this root symbol.

exchangestr
The exchange where this root symbol is traded.
equity_supplementary_mappings (pd.DataFrame, optional) – Additional mappings from values of abitrary type to assets.
chunk_size (int, optional) – The amount of rows to write to the SQLite table at once. This defaults to the default number of bind params in sqlite. If you have compiled sqlite3 with more bind or less params you may want to pass that value here.

See also

zipline.assets.asset_finder

write_direct(equities=None, equity_symbol_mappings=None, equity_supplementary_mappings=None, futures=None, exchanges=None, root_symbols=None, chunk_size=999)[source]#

Write asset metadata to a sqlite database in the format that it is stored in the assets db.

Parameters:

equities (pd.DataFrame, optional) –
The equity metadata. The columns for this dataframe are:

symbolstr
The ticker symbol for this equity.

asset_namestr
The full name for this asset.

start_datedatetime
The date when this asset was created.

end_datedatetime, optional
The last date we have trade data for this asset.

first_tradeddatetime, optional
The first date we have trade data for this asset.

auto_close_datedatetime, optional
The date on which to close any positions in this asset.

exchangestr
The exchange where this asset is traded.

The index of this dataframe should contain the sids.
futures (pd.DataFrame, optional) –
The future contract metadata. The columns for this dataframe are:

symbolstr
The ticker symbol for this futures contract.

root_symbolstr
The root symbol, or the symbol with the expiration stripped out.

asset_namestr
The full name for this asset.

start_datedatetime, optional
The date when this asset was created.

end_datedatetime, optional
The last date we have trade data for this asset.

first_tradeddatetime, optional
The first date we have trade data for this asset.

exchangestr
The exchange where this asset is traded.

notice_datedatetime
The date when the owner of the contract may be forced to take physical delivery of the contract’s asset.

expiration_datedatetime
The date when the contract expires.

auto_close_datedatetime
The date when the broker will automatically close any positions in this contract.

tick_sizefloat
The minimum price movement of the contract.

multiplier: float
The amount of the underlying asset represented by this contract.
exchanges (pd.DataFrame, optional) –
The exchanges where assets can be traded. The columns of this dataframe are:

exchangestr
The full name of the exchange.

canonical_namestr
The canonical name of the exchange.

country_codestr
The ISO 3166 alpha-2 country code of the exchange.
root_symbols (pd.DataFrame, optional) –
The root symbols for the futures contracts. The columns for this dataframe are:

root_symbolstr
The root symbol name.

root_symbol_idint
The unique id for this root symbol.

sectorstring, optional
The sector of this root symbol.

descriptionstring, optional
A short description of this root symbol.

exchangestr
The exchange where this root symbol is traded.
equity_supplementary_mappings (pd.DataFrame, optional) – Additional mappings from values of abitrary type to assets.
chunk_size (int, optional) – The amount of rows to write to the SQLite table at once. This defaults to the default number of bind params in sqlite. If you have compiled sqlite3 with more bind or less params you may want to pass that value here.

Readers#

class zipline.data.bcolz_daily_bars.BcolzDailyBarReader(table, read_all_threshold=3000)[source]#

Reader for raw pricing data written by BcolzDailyOHLCVWriter.

Parameters:

table (bcolz.ctable) – The ctable contaning the pricing data, with attrs corresponding to the Attributes list below.
read_all_threshold (int) – The number of equities at which; below, the data is read by reading a slice from the carray per asset. above, the data is read by pulling all of the data for all assets into memory and then indexing into that array for each day and asset pair. Used to tune performance of reads when using a small or large number of equities.

The table with which this loader interacts contains the following

attributes#

first_row#

Map from asset_id -> index of first row in the dataset with that id.

Type:: dict

last_row#

Map from asset_id -> index of last row in the dataset with that id.

Type:: dict

calendar_offset#

Map from asset_id -> calendar index of first row.

Type:: dict

start_session_ns#

Epoch ns of the first session used in this dataset.

Type:: int

end_session_ns#

Epoch ns of the last session used in this dataset.

Type:: int

calendar_name#

String identifier of trading calendar used (ie, “NYSE”).

Type:: str

We use first_row and last_row together to quickly find ranges of rows to

load when reading an asset's data into memory.

We use calendar_offset and calendar to orient loaded blocks within a

range of queried dates.

Notes

A Bcolz CTable is comprised of Columns and Attributes. The table with which this loader interacts contains the following columns:

[‘open’, ‘high’, ‘low’, ‘close’, ‘volume’, ‘day’, ‘id’].

The data in these columns is interpreted as follows:

Price columns (‘open’, ‘high’, ‘low’, ‘close’) are interpreted as 1000 * as-traded dollar value.
Volume is interpreted as as-traded volume.
Day is interpreted as seconds since midnight UTC, Jan 1, 1970.
Id is the asset id of the row.

The data in each column is grouped by asset and then sorted by day within each asset block.

The table is built to represent a long time range of data, e.g. ten years of equity data, so the lengths of each asset block is not equal to each other. The blocks are clipped to the known start and end date of each asset to cut down on the number of empty values that would need to be included to make a regular/cubic dataset.

When read across the open, high, low, close, and volume with the same index should represent the same asset and day.

currency_codes(sids)[source]#

Get currencies in which prices are quoted for the requested sids.

Assumes that a sid’s prices are always quoted in a single currency.

Parameters:: sids (np.array[int64]) – Array of sids for which currencies are needed.
Returns:: currency_codes – Array of currency codes for listing currencies of sids. Implementations should return None for sids whose currency is unknown.
Return type:: np.array[object]

get_last_traded_dt(asset, day)[source]#

Get the latest minute on or before dt in which asset traded.

If there are no trades on or before dt, returns pd.NaT.

Parameters:

asset (zipline.asset.Asset) – The asset for which to get the last traded minute.
dt (pd.Timestamp) – The minute at which to start searching for the last traded minute.

Returns:

last_traded – The dt of the last trade for the given asset, using the input dt as a vantage point.

Return type:

pd.Timestamp

get_value(sid, dt, field)[source]#

Parameters:

sid (int) – The asset identifier.
day (datetime64-like) – Midnight of the day for which data is requested.
colname (string) – The price field. e.g. (‘open’, ‘high’, ‘low’, ‘close’, ‘volume’)

Returns:

The spot price for colname of the given sid on the given day. Raises a NoDataOnDate exception if the given day and sid is before or after the date range of the equity. Returns -1 if the day is within the date range, but the price is 0.

Return type:

float

property last_available_dt#: returns: dt – The last session for which the reader can provide data. :rtype: pd.Timestamp

load_raw_arrays(columns, start_date, end_date, assets)[source]#

Parameters:

columns (list of str) – ‘open’, ‘high’, ‘low’, ‘close’, or ‘volume’
start_date (Timestamp) – Beginning of the window range.
end_date (Timestamp) – End of the window range.
assets (list of int) – The asset identifiers in the window.

Returns:

A list with an entry per field of ndarrays with shape (minutes in range, sids) with a dtype of float64, containing the values for the respective field over start and end dt range.

Return type:

list of np.ndarray

sid_day_index(sid, day)[source]#

Parameters:

sid (int) – The asset identifier.
day (datetime64-like) – Midnight of the day for which data is requested.

Returns:

Index into the data tape for the given sid and day. Raises a NoDataOnDate exception if the given day and sid is before or after the date range of the equity.

Return type:

int

class zipline.data.adjustments.SQLiteAdjustmentReader(conn)[source]#

Loads adjustments based on corporate actions from a SQLite database.

Expects data written in the format output by SQLiteAdjustmentWriter.

Parameters:: conn (str or sqlite3.Connection) – Connection from which to load data.

load_adjustments(dates, assets, should_include_splits, should_include_mergers, should_include_dividends, adjustment_type)[source]#

Load collection of Adjustment objects from underlying adjustments db.

Parameters:

dates (pd.DatetimeIndex) – Dates for which adjustments are needed.
assets (pd.Int64Index) – Assets for which adjustments are needed.
should_include_splits (bool) – Whether split adjustments should be included.
should_include_mergers (bool) – Whether merger adjustments should be included.
should_include_dividends (bool) – Whether dividend adjustments should be included.
adjustment_type (str) – Whether price adjustments, volume adjustments, or both, should be included in the output.

Returns:

adjustments – A dictionary containing price and/or volume adjustment mappings from index to adjustment objects to apply at that index.

Return type:

dict[str -> dict[int -> Adjustment]]

unpack_db_to_component_dfs(convert_dates=False)[source]#

Returns the set of known tables in the adjustments file in DataFrame form.

Parameters:: convert_dates (bool, optional) – By default, dates are returned in seconds since EPOCH. If convert_dates is True, all ints in date columns will be converted to datetimes.
Returns:: dfs – Dictionary which maps table name to the corresponding DataFrame version of the table, where all date columns have been coerced back from int to datetime.
Return type:: dict{str->DataFrame}

class zipline.assets.AssetFinder(engine, future_chain_predicates={'AD': functools.partial(<built-in function delivery_predicate>, {'Z', 'U', 'H', 'M'}), 'BP': functools.partial(<built-in function delivery_predicate>, {'Z', 'U', 'H', 'M'}), 'CD': functools.partial(<built-in function delivery_predicate>, {'Z', 'U', 'H', 'M'}), 'EL': functools.partial(<built-in function delivery_predicate>, {'Z', 'U', 'H', 'M'}), 'GC': functools.partial(<built-in function delivery_predicate>, {'M', 'Z', 'Q', 'V', 'G', 'J'}), 'JY': functools.partial(<built-in function delivery_predicate>, {'Z', 'U', 'H', 'M'}), 'ME': functools.partial(<built-in function delivery_predicate>, {'Z', 'U', 'H', 'M'}), 'PA': functools.partial(<built-in function delivery_predicate>, {'Z', 'U', 'H', 'M'}), 'PL': functools.partial(<built-in function delivery_predicate>, {'J', 'F', 'V', 'N'}), 'SV': functools.partial(<built-in function delivery_predicate>, {'H', 'N', 'Z', 'U', 'K'}), 'XG': functools.partial(<built-in function delivery_predicate>, {'M', 'Z', 'Q', 'V', 'G', 'J'}), 'YS': functools.partial(<built-in function delivery_predicate>, {'H', 'N', 'Z', 'U', 'K'})})[source]#

An AssetFinder is an interface to a database of Asset metadata written by an AssetDBWriter.

This class provides methods for looking up assets by unique integer id or by symbol. For historical reasons, we refer to these unique ids as ‘sids’.

Parameters:

engine (str or SQLAlchemy.engine) – An engine with a connection to the asset database to use, or a string that can be parsed by SQLAlchemy as a URI.
future_chain_predicates (dict) – A dict mapping future root symbol to a predicate function which accepts
be (a contract as a parameter and returns whether or not the contract should) –
chain. (included in the) –

See also

zipline.assets.AssetDBWriter

property equities_sids#: All of the sids for equities in the asset finder.

equities_sids_for_country_code(country_code)[source]#

Return all of the sids for a given country.

Parameters:: country_code (str) – An ISO 3166 alpha-2 country code.
Returns:: The sids whose exchanges are in this country.
Return type:: tuple[int]

equities_sids_for_exchange_name(exchange_name)[source]#

Return all of the sids for a given exchange_name.

Parameters:: exchange_name (str) –
Returns:: The sids whose exchanges are in this country.
Return type:: tuple[int]

property futures_sids#: All of the sids for futures consracts in the asset finder.

get_supplementary_field(sid, field_name, as_of_date)[source]#

Get the value of a supplementary field for an asset.

Parameters:

sid (int) – The sid of the asset to query.
field_name (str) – Name of the supplementary field.
as_of_date (pd.Timestamp, None) – The last known value on this date is returned. If None, a value is returned only if we’ve only ever had one value for this sid. If None and we’ve had multiple values, MultipleValuesFoundForSid is raised.

Raises:

NoValueForSid – If we have no values for this asset, or no values was known on this as_of_date.
MultipleValuesFoundForSid – If we have had multiple values for this asset over time, and None was passed for as_of_date.

group_by_type(sids)[source]#

Group a list of sids by asset type.

Parameters:: sids (list[int]) –
Returns:: types – A dict mapping unique asset types to lists of sids drawn from sids. If we fail to look up an asset, we assign it a key of None.
Return type:: dict[str or None -> list[int]]

lifetimes(dates, include_start_date, country_codes)[source]#

Compute a DataFrame representing asset lifetimes for the specified date range.

Parameters:

dates (pd.DatetimeIndex) – The dates for which to compute lifetimes.
include_start_date (bool) –
Whether or not to count the asset as alive on its start_date.

This is useful in a backtesting context where lifetimes is being used to signify “do I have data for this asset as of the morning of this date?” For many financial metrics, (e.g. daily close), data isn’t available for an asset until the end of the asset’s first day.
country_codes (iterable[str]) – The country codes to get lifetimes for.

Returns:

lifetimes – A frame of dtype bool with dates as index and an Int64Index of assets as columns. The value at lifetimes.loc[date, asset] will be True iff asset existed on date. If include_start_date is False, then lifetimes.loc[date, asset] will be false when date == asset.start_date.

Return type:

pd.DataFrame

See also

numpy.putmask, zipline.pipeline.engine.SimplePipelineEngine._compute_root_mask

lookup_asset_types(sids)[source]#

Retrieve asset types for a list of sids.

Parameters:: sids (list[int]) –
Returns:: types – Asset types for the provided sids.
Return type:: dict[sid -> str or None]

lookup_future_symbol(symbol)[source]#

Lookup a future contract by symbol.

Parameters:: symbol (str) – The symbol of the desired contract.
Returns:: future – The future contract referenced by symbol.
Return type:: Future
Raises:: SymbolNotFound – Raised when no contract named ‘symbol’ is found.

lookup_generic(obj, as_of_date, country_code)[source]#

Convert an object into an Asset or sequence of Assets.

This method exists primarily as a convenience for implementing user-facing APIs that can handle multiple kinds of input. It should not be used for internal code where we already know the expected types of our inputs.

Parameters:

obj (int, str, Asset, ContinuousFuture, or iterable) – The object to be converted into one or more Assets. Integers are interpreted as sids. Strings are interpreted as tickers. Assets and ContinuousFutures are returned unchanged.
as_of_date (pd.Timestamp or None) – Timestamp to use to disambiguate ticker lookups. Has the same semantics as in lookup_symbol.
country_code (str or None) – ISO-3166 country code to use to disambiguate ticker lookups. Has the same semantics as in lookup_symbol.

Returns:

matches, missing –

matches is the result of the conversion. missing is a list: containing any values that couldn’t be resolved. If obj is not an iterable, missing will be an empty list.

Return type:

tuple

lookup_symbol(symbol, as_of_date, fuzzy=False, country_code=None)[source]#

Lookup an equity by symbol.

Parameters:

symbol (str) – The ticker symbol to resolve.
as_of_date (datetime.datetime or None) – Look up the last owner of this symbol as of this datetime. If as_of_date is None, then this can only resolve the equity if exactly one equity has ever owned the ticker.
fuzzy (bool, optional) – Should fuzzy symbol matching be used? Fuzzy symbol matching attempts to resolve differences in representations for shareclasses. For example, some people may represent the A shareclass of BRK as BRK.A, where others could write BRK_A.
country_code (str or None, optional) – The country to limit searches to. If not provided, the search will span all countries which increases the likelihood of an ambiguous lookup.

Returns:

equity – The equity that held symbol on the given as_of_date, or the only equity to hold symbol if as_of_date is None.

Return type:

Equity

Raises:

SymbolNotFound – Raised when no equity has ever held the given symbol.
MultipleSymbolsFound – Raised when no as_of_date is given and more than one equity has held symbol. This is also raised when fuzzy=True and there are multiple candidates for the given symbol on the as_of_date. Also raised when no country_code is given and the symbol is ambiguous across multiple countries.

lookup_symbols(symbols, as_of_date, fuzzy=False, country_code=None)[source]#

Lookup a list of equities by symbol.

Equivalent to:

[finder.lookup_symbol(s, as_of, fuzzy) for s in symbols]

but potentially faster because repeated lookups are memoized.

Parameters:

symbols (sequence[str]) – Sequence of ticker symbols to resolve.
as_of_date (pd.Timestamp) – Forwarded to lookup_symbol.
fuzzy (bool, optional) – Forwarded to lookup_symbol.
country_code (str or None, optional) – The country to limit searches to. If not provided, the search will span all countries which increases the likelihood of an ambiguous lookup.

Returns:

equities

Return type:

list[Equity]

retrieve_all(sids, default_none=False)[source]#

Retrieve all assets in sids.

Parameters:

sids (iterable of int) – Assets to retrieve.
default_none (bool) – If True, return None for failed lookups. If False, raise SidsNotFound.

Returns:

assets – A list of the same length as sids containing Assets (or Nones) corresponding to the requested sids.

Return type:

list[Asset or None]

Raises:

SidsNotFound – When a requested sid is not found and default_none=False.

retrieve_asset(sid, default_none=False)[source]#: Retrieve the Asset for a given sid.

retrieve_equities(sids)[source]#

Retrieve Equity objects for a list of sids.

Users generally shouldn’t need to this method (instead, they should prefer the more general/friendly retrieve_assets), but it has a documented interface and tests because it’s used upstream.

Parameters:: sids (iterable[int]) –
Returns:: equities
Return type:: dict[int -> Equity]
Raises:: EquitiesNotFound – When any requested asset isn’t found.

retrieve_futures_contracts(sids)[source]#

Retrieve Future objects for an iterable of sids.

Users generally shouldn’t need to this method (instead, they should prefer the more general/friendly retrieve_assets), but it has a documented interface and tests because it’s used upstream.

Parameters:: sids (iterable[int]) –
Returns:: equities
Return type:: dict[int -> Equity]
Raises:: EquitiesNotFound – When any requested asset isn’t found.

property sids#: All the sids in the asset finder.

class zipline.data.data_portal.DataPortal(asset_finder, trading_calendar, first_trading_day, equity_daily_reader=None, equity_minute_reader=None, future_daily_reader=None, future_minute_reader=None, adjustment_reader=None, last_available_session=None, last_available_minute=None, minute_history_prefetch_length=1560, daily_history_prefetch_length=40)[source]#

Interface to all of the data that a zipline simulation needs.

This is used by the simulation runner to answer questions about the data, like getting the prices of assets on a given day or to service history calls.

Parameters:

asset_finder (zipline.assets.assets.AssetFinder) – The AssetFinder instance used to resolve assets.
trading_calendar (zipline.utils.calendar.exchange_calendar.TradingCalendar) – The calendar instance used to provide minute->session information.
first_trading_day (pd.Timestamp) – The first trading day for the simulation.
equity_daily_reader (BcolzDailyBarReader, optional) – The daily bar reader for equities. This will be used to service daily data backtests or daily history calls in a minute backetest. If a daily bar reader is not provided but a minute bar reader is, the minutes will be rolled up to serve the daily requests.
equity_minute_reader (BcolzMinuteBarReader, optional) – The minute bar reader for equities. This will be used to service minute data backtests or minute history calls. This can be used to serve daily calls if no daily bar reader is provided.
future_daily_reader (BcolzDailyBarReader, optional) – The daily bar ready for futures. This will be used to service daily data backtests or daily history calls in a minute backetest. If a daily bar reader is not provided but a minute bar reader is, the minutes will be rolled up to serve the daily requests.
future_minute_reader (BcolzFutureMinuteBarReader, optional) – The minute bar reader for futures. This will be used to service minute data backtests or minute history calls. This can be used to serve daily calls if no daily bar reader is provided.
adjustment_reader (SQLiteAdjustmentWriter, optional) – The adjustment reader. This is used to apply splits, dividends, and other adjustment data to the raw data from the readers.
last_available_session (pd.Timestamp, optional) – The last session to make available in session-level data.
last_available_minute (pd.Timestamp, optional) – The last minute to make available in minute-level data.

get_adjusted_value(asset, field, dt, perspective_dt, data_frequency, spot_value=None)[source]#

Returns a scalar value representing the value of the desired asset’s field at the given dt with adjustments applied.

Parameters:

asset (Asset) – The asset whose data is desired.
field ({'open', 'high', 'low', 'close', 'volume', 'price', 'last_traded'}) – The desired field of the asset.
dt (pd.Timestamp) – The timestamp for the desired value.
perspective_dt (pd.Timestamp) – The timestamp from which the data is being viewed back from.
data_frequency (str) – The frequency of the data to query; i.e. whether the data is ‘daily’ or ‘minute’ bars

Returns:

value – The value of the given field for asset at dt with any adjustments known by perspective_dt applied. The return type is based on the field requested. If the field is one of ‘open’, ‘high’, ‘low’, ‘close’, or ‘price’, the value will be a float. If the field is ‘volume’ the value will be a int. If the field is ‘last_traded’ the value will be a Timestamp.

Return type:

float, int, or pd.Timestamp

get_adjustments(assets, field, dt, perspective_dt)[source]#

Returns a list of adjustments between the dt and perspective_dt for the given field and list of assets

Parameters:

assets (list of type Asset, or Asset) – The asset, or assets whose adjustments are desired.
field ({'open', 'high', 'low', 'close', 'volume', 'price', 'last_traded'}) – The desired field of the asset.
dt (pd.Timestamp) – The timestamp for the desired value.
perspective_dt (pd.Timestamp) – The timestamp from which the data is being viewed back from.

Returns:

adjustments – The adjustments to that field.

Return type:

list[Adjustment]

get_current_future_chain(continuous_future, dt)[source]#

Retrieves the future chain for the contract at the given dt according the continuous_future specification.

Returns:: future_chain – A list of active futures, where the first index is the current contract specified by the continuous future definition, the second is the next upcoming contract and so on.
Return type:: list[Future]

get_fetcher_assets(dt)[source]#

Returns a list of assets for the current date, as defined by the fetcher data.

Returns:: list
Return type:: a list of Asset objects.

get_history_window(assets, end_dt, bar_count, frequency, field, data_frequency, ffill=True)[source]#

Public API method that returns a dataframe containing the requested history window. Data is fully adjusted.

Parameters:

assets (list of zipline.data.Asset objects) – The assets whose data is desired.
bar_count (int) – The number of bars desired.
frequency (string) – “1d” or “1m”
field (string) – The desired field of the asset.
data_frequency (string) – The frequency of the data to query; i.e. whether the data is ‘daily’ or ‘minute’ bars.
ffill (boolean) – Forward-fill missing values. Only has effect if field is ‘price’.

Return type:

A dataframe containing the requested data.

get_last_traded_dt(asset, dt, data_frequency)[source]#

Given an asset and dt, returns the last traded dt from the viewpoint of the given dt.

If there is a trade on the dt, the answer is dt provided.

get_scalar_asset_spot_value(asset, field, dt, data_frequency)[source]#

Public API method that returns a scalar value representing the value of the desired asset’s field at either the given dt.

Parameters:

assets (Asset) – The asset or assets whose data is desired. This cannot be an arbitrary AssetConvertible.
field ({'open', 'high', 'low', 'close', 'volume',) – ‘price’, ‘last_traded’} The desired field of the asset.
dt (pd.Timestamp) – The timestamp for the desired value.
data_frequency (str) – The frequency of the data to query; i.e. whether the data is ‘daily’ or ‘minute’ bars

Returns:

value – The spot value of field for asset The return type is based on the field requested. If the field is one of ‘open’, ‘high’, ‘low’, ‘close’, or ‘price’, the value will be a float. If the field is ‘volume’ the value will be a int. If the field is ‘last_traded’ the value will be a Timestamp.

Return type:

float, int, or pd.Timestamp

get_splits(assets, dt)[source]#

Returns any splits for the given sids and the given dt.

Parameters:

assets (container) – Assets for which we want splits.
dt (pd.Timestamp) – The date for which we are checking for splits. Note: this is expected to be midnight UTC.

Returns:

splits – List of splits, where each split is a (asset, ratio) tuple.

Return type:

list[(asset, float)]

get_spot_value(assets, field, dt, data_frequency)[source]#

Public API method that returns a scalar value representing the value of the desired asset’s field at either the given dt.

Parameters:

assets (Asset, ContinuousFuture, or iterable of same.) – The asset or assets whose data is desired.
field ({'open', 'high', 'low', 'close', 'volume',) – ‘price’, ‘last_traded’} The desired field of the asset.
dt (pd.Timestamp) – The timestamp for the desired value.
data_frequency (str) – The frequency of the data to query; i.e. whether the data is ‘daily’ or ‘minute’ bars

Returns:

Return type:

float, int, or pd.Timestamp

get_stock_dividends(sid, trading_days)[source]#

Returns all the stock dividends for a specific sid that occur in the given trading range.

Parameters:

sid (int) – The asset whose stock dividends should be returned.
trading_days (pd.DatetimeIndex) – The trading range.

Returns:

list (A list of objects with all relevant attributes populated.)
All timestamp fields are converted to pd.Timestamps.

handle_extra_source(source_df, sim_params)[source]#

Extra sources always have a sid column.

We expand the given data (by forward filling) to the full range of the simulation dates, so that lookup is fast during simulation.

class zipline.sources.benchmark_source.BenchmarkSource(benchmark_asset, trading_calendar, sessions, data_portal, emission_rate='daily', benchmark_returns=None)[source]#

daily_returns(start, end=None)[source]#

Returns the daily returns for the given period.

Parameters:

start (datetime) – The inclusive starting session label.
end (datetime, optional) – The inclusive ending session label. If not provided, treat start as a scalar key.

Returns:

returns – The returns in the given period. The index will be the trading calendar in the range [start, end]. If just start is provided, return the scalar value on that day.

Return type:

pd.Series or float

get_range(start_dt, end_dt)[source]#

Look up the returns for a given period.

Parameters:

start_dt (datetime) – The inclusive start label.
end_dt (datetime) – The inclusive end label.

Returns:

returns – The series of returns.

Return type:

pd.Series

: This method expects minute inputs if emission_rate == 'minute' and session labels when emission_rate == 'daily.

get_value(dt)[source]#

Look up the returns for a given dt.

Parameters:: dt (datetime) – The label to look up.
Returns:: returns – The returns at the given dt or session.
Return type:: float

: This method expects minute inputs if emission_rate == 'minute' and session labels when emission_rate == 'daily.

Bundles#

zipline.data.bundles.register(name='__no__default__', f='__no__default__', calendar_name='NYSE', start_session=None, end_session=None, minutes_per_day=390, create_writers=True)#

Parameters:

name (str) – The name of the bundle.
f (callable) –
The ingest function. This function will be passed:

environmapping
The environment this is being run with.

asset_db_writerAssetDBWriter
The asset db writer to write into.

minute_bar_writerBcolzMinuteBarWriter
The minute bar writer to write into.

daily_bar_writerBcolzDailyBarWriter
The daily bar writer to write into.

adjustment_writerSQLiteAdjustmentWriter
The adjustment db writer to write into.

calendartrading_calendars.TradingCalendar
The trading calendar to ingest for.

start_sessionpd.Timestamp
The first session of data to ingest.

end_sessionpd.Timestamp
The last session of data to ingest.

cacheDataFrameCache
A mapping object to temporarily store dataframes. This should be used to cache intermediates in case the load fails. This will be automatically cleaned up after a successful load.

show_progressbool
Show the progress for the current load where possible.
calendar_name (str, optional) – The name of a calendar used to align bundle data. Default is ‘NYSE’.
start_session (pd.Timestamp, optional) – The first session for which we want data. If not provided, or if the date lies outside the range supported by the calendar, the first_session of the calendar is used.
end_session (pd.Timestamp, optional) – The last session for which we want data. If not provided, or if the date lies outside the range supported by the calendar, the last_session of the calendar is used.
minutes_per_day (int, optional) – The number of minutes in each normal trading day.
create_writers (bool, optional) – Should the ingest machinery create the writers for the ingest function. This can be disabled as an optimization for cases where they are not needed, like the quantopian-quandl bundle.

Notes

This function my be used as a decorator, for example:

@register('quandl')
def quandl_ingest_function(...):
    ...

See also

zipline.data.bundles.bundles

zipline.data.bundles.ingest(name, environ=os.environ, date=None, show_progress=True)#

Ingest data for a given bundle.

Parameters:

name (str) – The name of the bundle.
environ (mapping, optional) – The environment variables. By default this is os.environ.
timestamp (datetime, optional) – The timestamp to use for the load. By default this is the current time.
assets_versions (Iterable[int], optional) – Versions of the assets db to which to downgrade.
show_progress (bool, optional) – Tell the ingest function to display the progress where possible.

zipline.data.bundles.load(name, environ=os.environ, date=None)#

Loads a previously ingested bundle.

Parameters:

name (str) – The name of the bundle.
environ (mapping, optional) – The environment variables. Defaults of os.environ.
timestamp (datetime, optional) – The timestamp of the data to lookup. Defaults to the current time.

Returns:

bundle_data – The raw data readers for this bundle.

Return type:

BundleData

zipline.data.bundles.unregister(name)#

Unregister a bundle.

Parameters:: name (str) – The name of the bundle to unregister.
Raises:: UnknownBundle – Raised when no bundle has been registered with the given name.

See also

zipline.data.bundles.bundles

zipline.data.bundles.bundles#: The bundles that have been registered as a mapping from bundle name to bundle data. This mapping is immutable and may only be updated through register() or unregister().

Risk Metrics#

Algorithm State#

class zipline.finance.ledger.Ledger(trading_sessions, capital_base, data_frequency)[source]#

The ledger tracks all orders and transactions as well as the current state of the portfolio and positions.

portfolio#

The updated portfolio being managed.

Type:: zipline.protocol.Portfolio

account#

The updated account being managed.

Type:: zipline.protocol.Account

position_tracker#

The current set of positions.

Type:: PositionTracker

todays_returns#

The current day’s returns. In minute emission mode, this is the partial day’s returns. In daily emission mode, this is daily_returns[session].

Type:: float

daily_returns_series#

The daily returns series. Days that have not yet finished will hold a value of np.nan.

Type:: pd.Series

daily_returns_array#

The daily returns as an ndarray. Days that have not yet finished will hold a value of np.nan.

Type:: np.ndarray

orders(dt=None)[source]#

Retrieve the dict-form of all of the orders in a given bar or for the whole simulation.

Parameters:: dt (pd.Timestamp or None, optional) – The particular datetime to look up order for. If not passed, or None is explicitly passed, all of the orders will be returned.
Returns:: orders – The order information.
Return type:: list[dict]

override_account_fields(settled_cash=sentinel('not_overridden'), accrued_interest=sentinel('not_overridden'), buying_power=sentinel('not_overridden'), equity_with_loan=sentinel('not_overridden'), total_positions_value=sentinel('not_overridden'), total_positions_exposure=sentinel('not_overridden'), regt_equity=sentinel('not_overridden'), regt_margin=sentinel('not_overridden'), initial_margin_requirement=sentinel('not_overridden'), maintenance_margin_requirement=sentinel('not_overridden'), available_funds=sentinel('not_overridden'), excess_liquidity=sentinel('not_overridden'), cushion=sentinel('not_overridden'), day_trades_remaining=sentinel('not_overridden'), leverage=sentinel('not_overridden'), net_leverage=sentinel('not_overridden'), net_liquidation=sentinel('not_overridden'))[source]#: Override fields on self.account.

property portfolio#

Compute the current portfolio.

Notes

This is cached, repeated access will not recompute the portfolio until the portfolio may have changed.

process_commission(commission)[source]#

Process the commission.

Parameters:: commission (zp.Event) – The commission being paid.

process_dividends(next_session, asset_finder, adjustment_reader)[source]#

Process dividends for the next session.

This will earn us any dividends whose ex-date is the next session as well as paying out any dividends whose pay-date is the next session

process_order(order)[source]#

Keep track of an order that was placed.

Parameters:: order (zp.Order) – The order to record.

process_splits(splits)[source]#

Processes a list of splits by modifying any positions as needed.

Parameters:: splits (list[(Asset, float)]) – A list of splits. Each split is a tuple of (asset, ratio).

process_transaction(transaction)[source]#

Add a transaction to ledger, updating the current state as needed.

Parameters:: transaction (zp.Transaction) – The transaction to execute.

transactions(dt=None)[source]#

Retrieve the dict-form of all of the transactions in a given bar or for the whole simulation.

Parameters:: dt (pd.Timestamp or None, optional) – The particular datetime to look up transactions for. If not passed, or None is explicitly passed, all of the transactions will be returned.
Returns:: transactions – The transaction information.
Return type:: list[dict]

update_portfolio()[source]#: Force a computation of the current portfolio state.

class zipline.protocol.Portfolio(start_date=None, capital_base=0.0)[source]#

Object providing read-only access to current portfolio state.

Parameters:

start_date (pd.Timestamp) – The start date for the period being recorded.
capital_base (float) – The starting value for the portfolio. This will be used as the starting cash, current cash, and portfolio value.

positions#

Dict-like object containing information about currently-held positions.

Type:: zipline.protocol.Positions

cash#

Amount of cash currently held in portfolio.

Type:: float

portfolio_value#

Current liquidation value of the portfolio’s holdings. This is equal to cash + sum(shares * price)

Type:: float

starting_cash#

Amount of cash in the portfolio at the start of the backtest.

Type:: float

property current_portfolio_weights#

Compute each asset’s weight in the portfolio by calculating its held value divided by the total value of all positions.

Each equity’s value is its price times the number of shares held. Each futures contract’s value is its unit price times number of shares held times the multiplier.

class zipline.protocol.Account[source]#: The account object tracks information about the trading account. The values are updated as the algorithm runs and its keys remain unchanged. If connected to a broker, one can update these values with the trading account values as reported by the broker.

class zipline.finance.ledger.PositionTracker(data_frequency)[source]#

The current state of the positions held.

Parameters:: data_frequency ({'daily', 'minute'}) – The data frequency of the simulation.

earn_dividends(cash_dividends, stock_dividends)[source]#

Given a list of dividends whose ex_dates are all the next trading day, calculate and store the cash and/or stock payments to be paid on each dividend’s pay date.

Parameters:

cash_dividends (iterable of (asset, amount, pay_date) namedtuples) –
stock_dividends (iterable of (asset, payment_asset, ratio, pay_date)) – namedtuples.

handle_splits(splits)[source]#

Processes a list of splits by modifying any positions as needed.

Parameters:: splits (list) – A list of splits. Each split is a tuple of (asset, ratio).
Returns:: int – position.
Return type:: The leftover cash from fractional shares after modifying each

pay_dividends(next_trading_day)[source]#: Returns a cash payment based on the dividends that should be paid out according to the accumulated bookkeeping of earned, unpaid, and stock dividends.

property stats#

The current status of the positions.

Returns:: stats – The current stats position stats.
Return type:: PositionStats

Notes

This is cached, repeated access will not recompute the stats until the stats may have changed.

class zipline.finance._finance_ext.PositionStats#

Computed values from the current positions.

gross_exposure#

The gross position exposure.

Type:: float64

gross_value#

The gross position value.

Type:: float64

long_exposure#

The exposure of just the long positions.

Type:: float64

long_value#

The value of just the long positions.

Type:: float64

net_exposure#

The net position exposure.

Type:: float64

net_value#

The net position value.

Type:: float64

short_exposure#

The exposure of just the short positions.

Type:: float64

short_value#

The value of just the short positions.

Type:: float64

longs_count#

The number of long positions.

Type:: int64

shorts_count#

The number of short positions.

Type:: int64

position_exposure_array#

The exposure of each position in the same order as position_tracker.positions.

Type:: np.ndarray[float64]

position_exposure_series#

The exposure of each position in the same order as position_tracker.positions. The index is the numeric sid of each asset.

Type:: pd.Series[float64]

Notes

position_exposure_array and position_exposure_series share the same underlying memory. The array interface should be preferred if you are doing access each minute for better performance.

position_exposure_array and position_exposure_series may be mutated when the position tracker next updates the stats. Do not rely on these objects being preserved across accesses to stats. If you need to freeze the values, you must take a copy.

Built-in Metrics#

class zipline.finance.metrics.metric.SimpleLedgerField(ledger_field, packet_field=None)[source]#

Emit the current value of a ledger field every bar or every session.

Parameters:

ledger_field (str) – The ledger field to read.
packet_field (str, optional) – The name of the field to populate in the packet. If not provided, ledger_field will be used.

class zipline.finance.metrics.metric.DailyLedgerField(ledger_field, packet_field=None)[source]#

Like SimpleLedgerField but also puts the current value in the cumulative_perf section.

Parameters:

ledger_field (str) – The ledger field to read.
packet_field (str, optional) – The name of the field to populate in the packet. If not provided, ledger_field will be used.

class zipline.finance.metrics.metric.StartOfPeriodLedgerField(ledger_field, packet_field=None)[source]#

Keep track of the value of a ledger field at the start of the period.

Parameters:

ledger_field (str) – The ledger field to read.
packet_field (str, optional) – The name of the field to populate in the packet. If not provided, ledger_field will be used.

class zipline.finance.metrics.metric.StartOfPeriodLedgerField(ledger_field, packet_field=None)[source]#

Keep track of the value of a ledger field at the start of the period.

Parameters:

ledger_field (str) – The ledger field to read.
packet_field (str, optional) – The name of the field to populate in the packet. If not provided, ledger_field will be used.

class zipline.finance.metrics.metric.Returns[source]#: Tracks the daily and cumulative returns of the algorithm.

class zipline.finance.metrics.metric.BenchmarkReturnsAndVolatility[source]#: Tracks daily and cumulative returns for the benchmark as well as the volatility of the benchmark returns.

class zipline.finance.metrics.metric.CashFlow[source]#

Tracks daily and cumulative cash flow.

Notes

For historical reasons, this field is named ‘capital_used’ in the packets.

class zipline.finance.metrics.metric.Orders[source]#: Tracks daily orders.

class zipline.finance.metrics.metric.Transactions[source]#: Tracks daily transactions.

class zipline.finance.metrics.metric.Positions[source]#: Tracks daily positions.

class zipline.finance.metrics.metric.ReturnsStatistic(function, field_name=None)[source]#

A metric that reports an end of simulation scalar or time series computed from the algorithm returns.

Parameters:

function (callable) – The function to call on the daily returns.
field_name (str, optional) – The name of the field. If not provided, it will be function.__name__.

class zipline.finance.metrics.metric.AlphaBeta[source]#: End of simulation alpha and beta to the benchmark.

class zipline.finance.metrics.metric.MaxLeverage[source]#: Tracks the maximum account leverage.

Metrics Sets#

zipline.finance.metrics.register(name, function=None)#

Parameters:

name (str) – The name of the metrics set
function (callable) – The callable which produces the metrics set.

Notes

This may be used as a decorator if only name is passed.

See also

zipline.finance.metrics.get_metrics_set, zipline.finance.metrics.unregister_metrics_set

zipline.finance.metrics.load(name)#

Return an instance of the metrics set registered with the given name.

Returns:: metrics – A new instance of the metrics set.
Return type:: set[Metric]
Raises:: ValueError – Raised when no metrics set is registered to name

zipline.finance.metrics.unregister(name)#

Unregister an existing metrics set.

Parameters:: name (str) – The name of the metrics set

See also

zipline.finance.metrics.register_metrics_set

zipline.data.finance.metrics.metrics_sets#: The metrics sets that have been registered as a mapping from metrics set name to load function. This mapping is immutable and may only be updated through register() or unregister().

Utilities#

Caching#

class zipline.utils.cache.CachedObject(value, expires)[source]#

A simple struct for maintaining a cached object with an expiration date.

Parameters:

value (object) – The object to cache.
expires (datetime-like) – Expiration date of value. The cache is considered invalid for dates strictly greater than expires.

Examples

>>> from pandas import Timestamp, Timedelta
>>> expires = Timestamp('2014', tz='UTC')
>>> obj = CachedObject(1, expires)
>>> obj.unwrap(expires - Timedelta('1 minute'))
1
>>> obj.unwrap(expires)
1
>>> obj.unwrap(expires + Timedelta('1 minute'))
... 
Traceback (most recent call last):
    ...
Expired: 2014-01-01 00:00:00+00:00

class zipline.utils.cache.ExpiringCache(cache=None, cleanup=<function ExpiringCache.<lambda>>)[source]#

A cache of multiple CachedObjects, which returns the wrapped the value or raises and deletes the CachedObject if the value has expired.

Parameters:

cache (dict-like, optional) – An instance of a dict-like object which needs to support at least: __del__, __getitem__, __setitem__ If None, than a dict is used as a default.
cleanup (callable, optional) – A method that takes a single argument, a cached object, and is called upon expiry of the cached object, prior to deleting the object. If not provided, defaults to a no-op.

Examples

>>> from pandas import Timestamp, Timedelta
>>> expires = Timestamp('2014', tz='UTC')
>>> value = 1
>>> cache = ExpiringCache()
>>> cache.set('foo', value, expires)
>>> cache.get('foo', expires - Timedelta('1 minute'))
1
>>> cache.get('foo', expires + Timedelta('1 minute'))
Traceback (most recent call last):
    ...
KeyError: 'foo'

class zipline.utils.cache.dataframe_cache(path=None, lock=None, clean_on_failure=True, serialization='pickle')[source]#

A disk-backed cache for dataframes.

dataframe_cache is a mutable mapping from string names to pandas DataFrame objects. This object may be used as a context manager to delete the cache directory on exit.

Parameters:

path (str, optional) – The directory path to the cache. Files will be written as path/<keyname>.
lock (Lock, optional) – Thread lock for multithreaded/multiprocessed access to the cache. If not provided no locking will be used.
clean_on_failure (bool, optional) – Should the directory be cleaned up if an exception is raised in the context manager.
serialize ({'msgpack', 'pickle:<n>'}, optional) – How should the data be serialized. If 'pickle' is passed, an optional pickle protocol can be passed like: 'pickle:3' which says to use pickle protocol 3.

Notes

The syntax cache[:] will load all key:value pairs into memory as a dictionary. The cache uses a temporary file format that is subject to change between versions of zipline.

class zipline.utils.cache.working_file(final_path, *args, **kwargs)[source]#

A context manager for managing a temporary file that will be moved to a non-temporary location if no exceptions are raised in the context.

Parameters:

final_path (str) – The location to move the file when committing.
*args – Forwarded to NamedTemporaryFile.
**kwargs – Forwarded to NamedTemporaryFile.

Notes

The file is moved on __exit__ if there are no exceptions. working_file uses shutil.move() to move the actual files, meaning it has as strong of guarantees as shutil.move().

class zipline.utils.cache.working_dir(final_path, *args, **kwargs)[source]#

A context manager for managing a temporary directory that will be moved to a non-temporary location if no exceptions are raised in the context.

Parameters:

final_path (str) – The location to move the file when committing.
*args – Forwarded to tmp_dir.
**kwargs – Forwarded to tmp_dir.

Notes

The file is moved on __exit__ if there are no exceptions. working_dir uses dir_util.copy_tree() to move the actual files, meaning it has as strong of guarantees as dir_util.copy_tree().

Command Line#

zipline.utils.cli.maybe_show_progress(it, show_progress, **kwargs)[source]#

Optionally show a progress bar for the given iterator.

Parameters:

it (iterable) – The underlying iterator.
show_progress (bool) – Should progress be shown.
**kwargs – Forwarded to the click progress bar.

Returns:

itercontext – A context manager whose enter is the actual iterator to use.

Return type:

context manager

Examples

with maybe_show_progress([1, 2, 3], True) as ns:
     for n in ns:
         ...