class WithInternationalDailyBarData(zf.WithAssetFinder):
"""
Fixture for generating international daily bars.
Eventually this should be moved into zipline.testing.fixtures and should
replace most of the existing machinery
"""
DAILY_BAR_START_DATE = zf.alias("START_DATE")
DAILY_BAR_END_DATE = zf.alias("END_DATE")
DAILY_BAR_LOOKBACK_DAYS = 0
INTERNATIONAL_PRICING_STARTING_PRICES = {
"XNYS": 100, # NYSE
"XTSE": 50, # Toronto Stock Exchange
"XLON": 25, # London Stock Exchange
}
# Assets in these countries will be quoted in one of the listed currencies.
INTERNATIONAL_PRICING_CURRENCIES = {
"XNYS": ["USD"],
"XTSE": ["CAD"],
"XLON": ["GBP", "EUR", "USD"],
}
assert (INTERNATIONAL_PRICING_STARTING_PRICES.keys() ==
INTERNATIONAL_PRICING_CURRENCIES.keys())
FX_RATES_CURRENCIES = ["USD", "CAD", "GBP", "EUR"]
@classmethod
def make_daily_bar_data(cls, assets, calendar, sessions):
# Generate prices corresponding to uniform random returns with a slight
# positive tendency.
start = cls.INTERNATIONAL_PRICING_STARTING_PRICES[calendar.name]
closes = random_tick_prices(start, len(sessions))
opens = closes - 0.05
highs = closes + 0.10
lows = closes - 0.10
volumes = np.arange(10000, 10000 + len(closes))
base_frame = pd.DataFrame(
{
"close": closes,
"open": opens,
"high": highs,
"low": lows,
"volume": volumes,
},
index=sessions,
)
for asset in assets:
sid = asset.sid
yield sid, base_frame + sid
@classmethod
def make_currency_codes(cls, calendar, assets):
currencies = cls.INTERNATIONAL_PRICING_CURRENCIES[calendar.name]
return pd.Series(index=assets,
data=list(islice(cycle(currencies), len(assets))))
@classmethod
def init_class_fixtures(cls):
super(WithInternationalDailyBarData, cls).init_class_fixtures()
cls.daily_bar_sessions = {}
cls.daily_bar_data = {}
cls.daily_bar_readers = {}
cls.daily_bar_currency_codes = {}
for (
calendar,
assets,
) in cls.assets_by_calendar.items():
name = calendar.name
start_delta = cls.DAILY_BAR_LOOKBACK_DAYS * calendar.day
start_session = cls.DAILY_BAR_START_DATE - start_delta
sessions = calendar.sessions_in_range(
start_session,
cls.DAILY_BAR_END_DATE,
)
cls.daily_bar_sessions[name] = sessions
cls.daily_bar_data[name] = dict(
cls.make_daily_bar_data(
assets=assets,
calendar=calendar,
sessions=sessions,
))
bar_data = cls.daily_bar_data[name]
df = (pd.concat(
bar_data, keys=bar_data.keys()).stack().unstack(0).swaplevel())
frames = {
field: frame.reset_index(level=0, drop=True)
for field, frame in df.groupby(level=0)
}
# panel = (pd.Panel.from_dict(cls.daily_bar_data[name])
# .transpose(2, 1, 0))
cls.daily_bar_currency_codes[name] = cls.make_currency_codes(
calendar,
assets,
)
cls.daily_bar_readers[name] = InMemoryDailyBarReader(
frames=frames,
calendar=calendar,
currency_codes=cls.daily_bar_currency_codes[name],
)
class TestDataPortalExplicitLastAvailable(DataPortalTestBase, ZiplineTestCase):
DATA_PORTAL_LAST_AVAILABLE_SESSION = alias("START_DATE")
DATA_PORTAL_LAST_AVAILABLE_MINUTE = alias("END_DATE")
class TestDataPortalExplicitLastAvailable(DataPortalTestBase):
DATA_PORTAL_LAST_AVAILABLE_SESSION = alias('START_DATE')
DATA_PORTAL_LAST_AVAILABLE_MINUTE = alias('END_DATE')
class WithInternationalDailyBarData(zf.WithAssetFinder):
"""
Fixture for generating international daily bars.
Eventually this should be moved into zipline.testing.fixtures and should
replace most of the existing machinery
"""
DAILY_BAR_START_DATE = zf.alias('START_DATE')
DAILY_BAR_END_DATE = zf.alias('END_DATE')
DAILY_BAR_LOOKBACK_DAYS = 0
INTERNATIONAL_PRICING_STARTING_PRICES = {
'XNYS': 100, # NYSE
'XTSE': 50, # Toronto Stock Exchange
'XLON': 25, # London Stock Exchange
}
# Assets in these countries will be quoted in one of the listed currencies.
INTERNATIONAL_PRICING_CURRENCIES = {
'XNYS': ['USD'],
'XTSE': ['CAD'],
'XLON': ['GBP', 'EUR', 'USD'],
}
assert (INTERNATIONAL_PRICING_STARTING_PRICES.keys() ==
INTERNATIONAL_PRICING_CURRENCIES.keys())
FX_RATES_CURRENCIES = ["USD", "CAD", "GBP", "EUR"]
@classmethod
def make_daily_bar_data(cls, assets, calendar, sessions):
# Generate prices corresponding to uniform random returns with a slight
# positive tendency.
start = cls.INTERNATIONAL_PRICING_STARTING_PRICES[calendar.name]
closes = random_tick_prices(start, len(sessions))
opens = closes - 0.05
highs = closes + 0.10
lows = closes - 0.10
volumes = np.arange(10000, 10000 + len(closes))
base_frame = pd.DataFrame(
{
'close': closes,
'open': opens,
'high': highs,
'low': lows,
'volume': volumes,
},
index=sessions)
for asset in assets:
sid = asset.sid
yield sid, base_frame + sid
@classmethod
def make_currency_codes(cls, calendar, assets):
currencies = cls.INTERNATIONAL_PRICING_CURRENCIES[calendar.name]
return pd.Series(index=assets,
data=list(islice(cycle(currencies), len(assets))))
@classmethod
def init_class_fixtures(cls):
super(WithInternationalDailyBarData, cls).init_class_fixtures()
cls.daily_bar_sessions = {}
cls.daily_bar_data = {}
cls.daily_bar_readers = {}
cls.daily_bar_currency_codes = {}
for calendar, assets, in cls.assets_by_calendar.items():
name = calendar.name
start_delta = cls.DAILY_BAR_LOOKBACK_DAYS * calendar.day
start_session = cls.DAILY_BAR_START_DATE - start_delta
sessions = calendar.sessions_in_range(
start_session,
cls.DAILY_BAR_END_DATE,
)
cls.daily_bar_sessions[name] = sessions
cls.daily_bar_data[name] = dict(
cls.make_daily_bar_data(
assets=assets,
calendar=calendar,
sessions=sessions,
))
panel = pd.concat(cls.daily_bar_data[name], axis=0)
dict_data = {}
for column in panel.columns:
dict_data[column] = pd.DataFrame(
panel[column]).unstack(level=0)
dict_data[column].columns = dict_data[
column].columns.droplevel(0)
cls.daily_bar_currency_codes[name] = cls.make_currency_codes(
calendar,
assets,
)
cls.daily_bar_readers[name] = InMemoryDailyBarReader.from_panel(
dict_data,
calendar,
currency_codes=cls.daily_bar_currency_codes[name],
)
class WithInternationalDailyBarData(zf.WithAssetFinder):
"""
Fixture for generating international daily bars.
Eventually this should be moved into zipline.testing.fixtures and should
replace most of the existing machinery
"""
DAILY_BAR_START_DATE = zf.alias('START_DATE')
DAILY_BAR_END_DATE = zf.alias('END_DATE')
DAILY_BAR_LOOKBACK_DAYS = 0
INTERNATIONAL_PRICING_STARTING_PRICES = {
'NYSE': 100,
'TSX': 50,
'LSE': 25,
}
@classmethod
def make_daily_bar_data(cls, assets, calendar, sessions):
# Generate prices corresponding to uniform random returns with a slight
# positive tendency.
start = cls.INTERNATIONAL_PRICING_STARTING_PRICES[calendar.name]
closes = random_tick_prices(start, len(sessions))
opens = closes - 0.05
highs = closes + 0.10
lows = closes - 0.10
volumes = np.arange(10000, 10000 + len(closes))
base_frame = pd.DataFrame(
{
'close': closes,
'open': opens,
'high': highs,
'low': lows,
'volume': volumes,
},
index=sessions)
for asset in assets:
sid = asset.sid
yield sid, base_frame + sid
@classmethod
def init_class_fixtures(cls):
super(WithInternationalDailyBarData, cls).init_class_fixtures()
cls.daily_bar_sessions = {}
cls.daily_bar_data = {}
cls.daily_bar_readers = {}
for calendar, assets, in cls.assets_by_calendar.items():
name = calendar.name
start_delta = cls.DAILY_BAR_LOOKBACK_DAYS * calendar.day
start_session = cls.DAILY_BAR_START_DATE - start_delta
sessions = calendar.sessions_in_range(
start_session,
cls.DAILY_BAR_END_DATE,
)
cls.daily_bar_sessions[name] = sessions
cls.daily_bar_data[name] = dict(
cls.make_daily_bar_data(
assets=assets,
calendar=calendar,
sessions=sessions,
))
panel = (pd.Panel.from_dict(cls.daily_bar_data[name]).transpose(
2, 1, 0))
cls.daily_bar_readers[name] = InMemoryDailyBarReader.from_panel(
panel,
calendar,
)