def __init__(self, dataHandler,
events,
startDate,
assets,
initialCapital=100000.0,
benchmark=None,
portfolioType=PortfolioType.CashManageable):
self.dataHandler = dataHandler
self.events = events
self.allTradableAssets = self.dataHandler.allTradableAssets
self.startDate = startDate
self.initialCapital = initialCapital
self.benchmark = benchmark
self.assets = assets
self.positionsBook = StocksPositionsBook(assets)
self.portfolioType = portfolioType
self.allPositions = self.constructAllPositions()
self.currentPosition = defaultdict(int, [(s, 0) for s in self.allTradableAssets])
self.allHoldings = [] # self.constructAllHoldings()
self.currentHoldings = self.constructCurrentHoldings()
self.orderBook = None
vp_settings.set_source(Settings.data_source)
def __init__(self, dataHandler,
events,
startDate,
assets,
initialCapital=100000.0,
benchmark=None,
portfolioType=PortfolioType.CashManageable):
self.dataHandler = dataHandler
self.events = events
self.allTradableAssets = self.dataHandler.allTradableAssets
self.startDate = startDate
self.initialCapital = initialCapital
self.benchmark = benchmark
self.assets = assets
self.positionsBook = StocksPositionsBook(assets)
self.portfolioType = portfolioType
self.allPositions = self.constructAllPositions()
self.currentPosition = defaultdict(int, [(s, 0) for s in self.allTradableAssets])
self.allHoldings = [] # self.constructAllHoldings()
self.currentHoldings = self.constructCurrentHoldings()
self.orderBook = None
vp_settings.set_source(Settings.data_source)
def __init__(self, dataHandler,
events,
startDate,
assets,
initialCapital=100000.0,
benchmark=None,
portfolioType=PortfolioType.CashManageable):
self.dataHandler = dataHandler
self.events = events
self.tradableAssets = self.dataHandler.tradableAssets
self.startDate = startDate
self.initialCapital = initialCapital
self.benchmark = benchmark
self.assets = assets
self.positionsBook = StocksPositionsBook(assets)
self.portfolioType = portfolioType
self.allPositions = self.constructAllPositions()
self.currentPosition = dict((s, 0) for s in self.tradableAssets)
self.allHoldings = self.constructAllHoldings()
self.currentHoldings = self.constructCurrentHoldings()
class Portfolio(object):
def __init__(self, dataHandler,
events,
startDate,
assets,
initialCapital=100000.0,
benchmark=None,
portfolioType=PortfolioType.CashManageable):
self.dataHandler = dataHandler
self.events = events
self.allTradableAssets = self.dataHandler.allTradableAssets
self.startDate = startDate
self.initialCapital = initialCapital
self.benchmark = benchmark
self.assets = assets
self.positionsBook = StocksPositionsBook(assets)
self.portfolioType = portfolioType
self.allPositions = self.constructAllPositions()
self.currentPosition = defaultdict(int, [(s, 0) for s in self.allTradableAssets])
self.allHoldings = [] # self.constructAllHoldings()
self.currentHoldings = self.constructCurrentHoldings()
self.orderBook = None
vp_settings.set_source(Settings.data_source)
def constructAllPositions(self):
d = dict((k, v) for k, v in [(s, 0) for s in self.allTradableAssets])
d['datetime'] = self.startDate
return [d]
def constructAllHoldings(self):
d = dict((k, v) for k, v in [(s, 0) for s in self.allTradableAssets])
d['datetime'] = self.startDate
d['cash'] = self.initialCapital
d['margin'] = 0.0
d['commission'] = 0.0
d['pnl'] = 0.
d['total'] = self.initialCapital
return [d]
def constructCurrentHoldings(self):
d = dict((k, v) for k, v in [(s, 0) for s in self.allTradableAssets])
d['datetime'] = self.startDate
d['cash'] = self.initialCapital
d['margin'] = 0.0
d['commission'] = 0.0
d['pnl'] = 0.
d['total'] = self.initialCapital
return d
def updateTimeindex(self):
latestDatetime = self.dataHandler.currentTimeIndex
dh = dict((s, 0) for s in self.allTradableAssets)
dh['datetime'] = latestDatetime
dh['cash'] = self.currentHoldings['cash']
dh['commission'] = self.currentHoldings['commission']
dh['margin'] = 0.
dh['total'] = self.currentHoldings['total']
dh['pnl'] = self.currentHoldings['pnl']
for s in self.allTradableAssets:
bookValue = 0.
bookPnL = 0.
margin = 0.
if self.currentPosition[s]:
currentCost = self.dataHandler.getLatestBarValue(s, 'close') * self.assets[s].multiplier
bookValue, bookPnL, margin = self.positionsBook.getBookValueAndBookPnL(s, currentCost)
dh[s] = bookValue
dh['margin'] += margin
dh['pnl'] += bookPnL
dh['total'] += bookPnL
self.allHoldings.append(dh)
def updatePositionFromFill(self, fill):
fillDir = fill.direction
if fillDir == OrderDirection.BUY or fillDir == OrderDirection.BUY_BACK:
self.currentPosition[fill.symbol] += fill.quantity
else:
self.currentPosition[fill.symbol] -= fill.quantity
def updateHoldingsFromFill(self, fill, pnl):
self.currentHoldings[fill.symbol] += fill.fillCost
self.currentHoldings['commission'] += fill.commission
if not isClose(fill.fillCost, 0.):
self.currentHoldings['cash'] -= (fill.fillCost + fill.commission)
else:
self.currentHoldings['cash'] += (pnl - fill.commission)
self.currentHoldings['pnl'] += pnl - fill.commission
self.currentHoldings['total'] += pnl - fill.commission
fillDir = fill.direction
if fillDir == OrderDirection.BUY_BACK or fillDir == OrderDirection.SELL_SHORT:
self.currentHoldings['margin'] -= fill.fillCost
def updateFill(self, event):
posClosed, posOpen, pnl = self.positionsBook.updatePositionsByFill(event)
self.updatePositionFromFill(event)
self.updateHoldingsFromFill(event, pnl)
self.filledBook.updateFromFillEvent(event)
def cancelOrders(self, timeIndex, posBook):
self.orderBook.cancelOrders(timeIndex, posBook)
def generateNaiveOrder(self, signal):
order = None
symbol = signal.symbol
direction = signal.signalType
mktQuantity = signal.quantity
curQuantity = self.currentPosition[symbol]
orderType = 'MKT'
if direction == 'LONG' and curQuantity == 0:
order = OrderEvent(symbol, orderType, mktQuantity, 1)
if direction == 'SHORT' and curQuantity == 0:
order = OrderEvent(symbol, orderType, mktQuantity, -1)
if direction == 'EXIT' and curQuantity > 0:
order = OrderEvent(symbol, orderType, abs(curQuantity), -1)
if direction == 'EXIT' and curQuantity < 0:
order = OrderEvent(symbol, orderType, abs(curQuantity), 1)
return order
def updateSignal(self, event):
if event.type == 'SIGNAL':
orderEvent = self.generateNaiveOrder(event)
self.events.put(orderEvent)
def _createFullNotionalEquityCurve(self, curve):
rawpos = curve.drop(['cash', 'commission', 'total', 'margin', 'pnl'], axis=1)
notionals = rawpos.abs().sum(axis=1).fillna(0)
notional_directions = rawpos.sum(axis=1).fillna(0).apply(sign)
pnldiffs = curve['pnl'].diff()
# to handl the case when pnl is shown in null notional day
cumPnL = 0.
returns = []
for notional, pnldiff, direction in zip(notionals, pnldiffs, notional_directions):
if not isClose(notional):
r = direction * np.log(direction * notional / (direction * notional - pnldiff - cumPnL))
returns.append(r)
cumPnL = 0.
elif not np.isnan(pnldiff):
returns.append(0.)
cumPnL += pnldiff
else:
returns.append(0.)
curve['return'] = returns
curve.fillna(0., inplace=True)
curve['equity_curve'] = np.exp(curve['return'].cumsum())
def createEquityCurveDataframe(self):
curve = pd.DataFrame(self.allHoldings)
curve.set_index('datetime', inplace=True)
if self.portfolioType == PortfolioType.FullNotional:
self._createFullNotionalEquityCurve(curve)
else:
curve['return'] = np.log(curve['total'] / curve['total'].shift(1))
curve.dropna(inplace=True)
curve['equity_curve'] = np.exp(curve['return'].cumsum())
self.equityCurve = curve.dropna()
def outputSummaryStats(self, curve, other_curves, plot):
returns = curve['return']
if hasattr(self.dataHandler, "benchmarkData"):
benchmarkReturns = self.dataHandler.benchmarkData['return']
benchmarkReturns.name = self.benchmark
else:
benchmarkReturns = None
perf_metric, perf_df, rollingRisk = createPerformanceTearSheet(returns=returns,
benchmarkReturns=benchmarkReturns,
other_curves=other_curves,
plot=plot)
if self.portfolioType == PortfolioType.FullNotional:
positons = curve.drop(['cash', 'commission', 'total', 'return', 'margin', 'equity_curve', 'pnl'], axis=1)
else:
positons = curve.drop(['commission', 'total', 'return', 'margin', 'equity_curve', 'pnl'], axis=1)
aggregated_positons = createPostionTearSheet(positons, plot=plot)
transactions = extractTransactionFromFilledBook(self.filledBook.view())
if hasattr(self.dataHandler, "_freq"):
if self.dataHandler._freq == 0:
freq = 'M'
else:
freq = 'D'
turnover_rate = createTranscationTearSheet(transactions, positons, freq=freq, plot=plot)
else:
turnover_rate = createTranscationTearSheet(transactions, positons, plot=plot)
if plot:
plt.show()
return perf_metric, perf_df, rollingRisk, aggregated_positons, transactions, turnover_rate
class Portfolio(object):
def __init__(self, dataHandler,
events,
startDate,
assets,
initialCapital=100000.0,
benchmark=None,
portfolioType=PortfolioType.CashManageable):
self.dataHandler = dataHandler
self.events = events
self.tradableAssets = self.dataHandler.tradableAssets
self.startDate = startDate
self.initialCapital = initialCapital
self.benchmark = benchmark
self.assets = assets
self.positionsBook = StocksPositionsBook(assets)
self.portfolioType = portfolioType
self.allPositions = self.constructAllPositions()
self.currentPosition = defaultdict(int, [(s, 0) for s in self.tradableAssets])
self.allHoldings = self.constructAllHoldings()
self.currentHoldings = self.constructCurrentHoldings()
vp_settings.set_source(Settings.data_source)
def constructAllPositions(self):
d = dict((k, v) for k, v in [(s, 0) for s in self.tradableAssets])
d['datetime'] = self.startDate
return [d]
def constructAllHoldings(self):
d = dict((k, v) for k, v in [(s, 0) for s in self.tradableAssets])
d['datetime'] = self.startDate
d['cash'] = self.initialCapital
d['margin'] = 0.0
d['commission'] = 0.0
d['pnl'] = 0.
d['total'] = self.initialCapital
return [d]
def constructCurrentHoldings(self):
d = dict((k, v) for k, v in [(s, 0) for s in self.tradableAssets])
d['datetime'] = self.startDate
d['cash'] = self.initialCapital
d['margin'] = 0.0
d['commission'] = 0.0
d['pnl'] = 0.
d['total'] = self.initialCapital
return d
def updateTimeindex(self):
latestDatetime = self.dataHandler.currentTimeIndex
dh = dict((s, 0) for s in self.tradableAssets)
dh['datetime'] = latestDatetime
dh['cash'] = self.currentHoldings['cash']
dh['commission'] = self.currentHoldings['commission']
dh['margin'] = self.currentHoldings['margin']
dh['total'] = self.currentHoldings['total']
dh['pnl'] = self.currentHoldings['pnl']
for s in self.tradableAssets:
bookValue = 0.
bookPnL = 0.
if self.currentPosition[s]:
currentCost = self.dataHandler.getLatestBarValue(s, 'close') * self.assets[s].multiplier
bookValue, bookPnL = self.positionsBook.getBookValueAndBookPnL(s, currentCost)
dh[s] = bookValue
dh['pnl'] += bookPnL
dh['total'] += bookPnL
self.allHoldings.append(dh)
def updatePositionFromFill(self, fill):
fillDir = fill.direction
self.currentPosition[fill.symbol] += fillDir * fill.quantity
def updateHoldingsFromFill(self, fill, pnl):
self.currentHoldings[fill.symbol] += fill.fillCost
self.currentHoldings['commission'] += fill.commission
if not isClose(fill.fillCost, 0.):
self.currentHoldings['cash'] -= (fill.fillCost + fill.commission)
else:
self.currentHoldings['cash'] += (pnl - fill.commission)
self.currentHoldings['pnl'] += pnl - fill.commission
self.currentHoldings['total'] += pnl - fill.commission
def updateFill(self, event):
posClosed, posOpen, pnl = self.positionsBook.updatePositionsByFill(event)
self.updatePositionFromFill(event)
self.updateHoldingsFromFill(event, pnl)
self.filledBook.updateFromFillEvent(event)
def generateNaiveOrder(self, signal):
order = None
symbol = signal.symbol
direction = signal.signalType
mktQuantity = signal.quantity
curQuantity = self.currentPosition[symbol]
orderType = 'MKT'
if direction == 'LONG' and curQuantity == 0:
order = OrderEvent(symbol, orderType, mktQuantity, 1)
if direction == 'SHORT' and curQuantity == 0:
order = OrderEvent(symbol, orderType, mktQuantity, -1)
if direction == 'EXIT' and curQuantity > 0:
order = OrderEvent(symbol, orderType, abs(curQuantity), -1)
if direction == 'EXIT' and curQuantity < 0:
order = OrderEvent(symbol, orderType, abs(curQuantity), 1)
return order
def updateSignal(self, event):
if event.type == 'SIGNAL':
orderEvent = self.generateNaiveOrder(event)
self.events.put(orderEvent)
def _createFullNotionalEquityCurve(self, curve):
rawpos = curve.drop(['cash', 'commission', 'total', 'margin', 'pnl'], axis=1)
notionals = rawpos.abs().sum(axis=1).fillna(0)
notional_directions = rawpos.sum(axis=1).fillna(0).apply(sign)
pnldiffs = curve['pnl'].diff()
# to handl the case when pnl is shown in null notional day
cumPnL = 0.
returns = []
for notional, pnldiff, direction in zip(notionals, pnldiffs, notional_directions):
if not isClose(notional):
r = direction * np.log(direction * notional / (direction * notional - pnldiff - cumPnL))
returns.append(r)
cumPnL = 0.
elif not np.isnan(pnldiff):
returns.append(0.)
cumPnL += pnldiff
else:
returns.append(0.)
curve['return'] = returns
curve.fillna(0., inplace=True)
curve['equity_curve'] = np.exp(curve['return'].cumsum())
def createEquityCurveDataframe(self):
curve = pd.DataFrame(self.allHoldings)
curve.set_index('datetime', inplace=True)
if self.portfolioType == PortfolioType.FullNotional:
self._createFullNotionalEquityCurve(curve)
else:
curve['return'] = np.log(curve['total'] / curve['total'].shift(1))
curve.dropna(inplace=True)
curve['equity_curve'] = np.exp(curve['return'].cumsum())
self.equityCurve = curve.dropna()
def outputSummaryStats(self, curve, plot):
returns = curve['return']
if hasattr(self.dataHandler, "benchmarkData"):
benchmarkReturns = self.dataHandler.benchmarkData['return']
benchmarkReturns.name = self.benchmark
else:
benchmarkReturns = None
perf_metric, perf_df, rollingRisk = createPerformanceTearSheet(returns=returns, benchmarkReturns=benchmarkReturns, plot=plot)
if self.portfolioType == PortfolioType.FullNotional:
positons = curve.drop(['cash', 'commission', 'total', 'return', 'margin', 'equity_curve', 'pnl'], axis=1)
else:
positons = curve.drop(['commission', 'total', 'return', 'margin', 'equity_curve', 'pnl'], axis=1)
aggregated_positons = createPostionTearSheet(positons, plot=plot)
transactions = extractTransactionFromFilledBook(self.filledBook.view())
turnover_rate = createTranscationTearSheet(transactions, positons, plot=plot)
if plot:
plt.show()
return perf_metric, perf_df, rollingRisk, aggregated_positons, transactions, turnover_rate
def setUp(self):
assets = {'600000': XSHGStock, 'IF1501': IndexFutures}
self.positionsBook = StocksPositionsBook(assets)
class TestPositionsBook(unittest.TestCase):
def setUp(self):
assets = {'600000': XSHGStock, 'IF1501': IndexFutures}
self.positionsBook = StocksPositionsBook(assets)
def testUpdatedFromOrder(self):
self.positionsBook.updatePositionsByOrder('600000',
dt.date(2015, 11, 16), 100,
1)
self.assertEqual(self.positionsBook._allPositions, {})
def testUpdatedFromFillEvent(self):
self.positionsBook.updatePositionsByFill('600000',
dt.date(2015, 11, 16), 100, 1)
self.assertEqual(self.positionsBook._allPositions['600000'],
([dt.date(2015, 11, 16)], [100], [0], [1]))
def testUpdatedFromOrderWithNonEmptyPositionBook(self):
self.positionsBook.updatePositionsByFill('600000',
dt.date(2015, 11, 13), 300, 1)
self.assertEqual(
self.positionsBook.avaliableForTrade('600000',
dt.date(2015, 11, 16)),
(300, 0))
self.positionsBook.updatePositionsByOrder('600000',
dt.date(2015, 11, 16), 100,
-1)
self.assertEqual(self.positionsBook._allPositions['600000'],
([dt.date(2015, 11, 13)], [300], [100], [1]))
self.assertEqual(
self.positionsBook.avaliableForTrade('600000',
dt.date(2015, 11, 16)),
(200, 0))
def testUpdatedSellShortOrderWithShortForbiddenAsset(self):
with self.assertRaises(ValueError):
self.positionsBook.updatePositionsByOrder('600000',
dt.date(2015, 11, 16),
100, -1)
def testUpdatedSellShortOrderWithShortAllowedAsset(self):
self.positionsBook.updatePositionsByOrder('IF1501',
dt.date(2015, 11, 16), 100,
-1)
def testUpdatedFromFillWithBothLongAndShot(self):
self.positionsBook.updatePositionsByOrder('IF1501',
dt.date(2015, 11, 13), 1, -1)
self.positionsBook.updatePositionsByFill('IF1501',
dt.date(2015, 11, 13), 1, -1)
self.assertEqual(
self.positionsBook.avaliableForTrade('IF1501',
dt.date(2015, 11, 13)),
(0, 1))
self.positionsBook.updatePositionsByOrder('IF1501',
dt.date(2015, 11, 16), 2, 1)
self.positionsBook.updatePositionsByFill('IF1501',
dt.date(2015, 11, 16), 1, 1)
self.positionsBook.updatePositionsByFill('IF1501',
dt.date(2015, 11, 16), 1, 1)
self.assertEqual(self.positionsBook._allPositions['IF1501'],
([dt.date(2015, 11, 16)], [1], [0], [1]))
def setUp(self):
assets = {'600000': XSHGStock,
'IF1501': IndexFutures}
self.positionsBook = StocksPositionsBook(assets)
class TestPositionsBook(unittest.TestCase):
def setUp(self):
assets = {'600000': XSHGStock,
'IF1501': IndexFutures}
self.positionsBook = StocksPositionsBook(assets)
def testUpdatedFromOrder(self):
self.positionsBook.updatePositionsByOrder('600000', dt.date(2015, 11, 16), 100, 1)
self.assertEqual(self.positionsBook._allPositions, {})
def testUpdatedFromFillEvent(self):
self.positionsBook.updatePositionsByFill('600000', dt.date(2015, 11, 16), 100, 1)
self.assertEqual(self.positionsBook._allPositions['600000'], ([dt.date(2015, 11, 16)],
[100],
[0],
[1]))
def testUpdatedFromOrderWithNonEmptyPositionBook(self):
self.positionsBook.updatePositionsByFill('600000', dt.date(2015, 11, 13), 300, 1)
self.assertEqual(self.positionsBook.avaliableForTrade('600000', dt.date(2015, 11, 16)), (300, 0))
self.positionsBook.updatePositionsByOrder('600000', dt.date(2015, 11, 16), 100, -1)
self.assertEqual(self.positionsBook._allPositions['600000'], ([dt.date(2015, 11, 13)],
[300],
[100],
[1]))
self.assertEqual(self.positionsBook.avaliableForTrade('600000', dt.date(2015, 11, 16)), (200, 0))
def testUpdatedSellShortOrderWithShortForbiddenAsset(self):
with self.assertRaises(ValueError):
self.positionsBook.updatePositionsByOrder('600000', dt.date(2015, 11, 16), 100, -1)
def testUpdatedSellShortOrderWithShortAllowedAsset(self):
self.positionsBook.updatePositionsByOrder('IF1501', dt.date(2015, 11, 16), 100, -1)
def testUpdatedFromFillWithBothLongAndShot(self):
self.positionsBook.updatePositionsByOrder('IF1501', dt.date(2015, 11, 13), 1, -1)
self.positionsBook.updatePositionsByFill('IF1501', dt.date(2015, 11, 13), 1, -1)
self.assertEqual(self.positionsBook.avaliableForTrade('IF1501', dt.date(2015, 11, 13)), (0, 1))
self.positionsBook.updatePositionsByOrder('IF1501', dt.date(2015, 11, 16), 2, 1)
self.positionsBook.updatePositionsByFill('IF1501', dt.date(2015, 11, 16), 1, 1)
self.positionsBook.updatePositionsByFill('IF1501', dt.date(2015, 11, 16), 1, 1)
self.assertEqual(self.positionsBook._allPositions['IF1501'], ([dt.date(2015, 11, 16)],
[1],
[0],
[1]))
