def test_write_one_ohlcv_with_ratios(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={"open": [10.0], "high": [20.0], "low": [30.0], "close": [40.0], "volume": [50.0]}, index=[minute]
)
# Create a new writer with `ohlc_ratios_per_sid` defined.
writer_with_ratios = BcolzMinuteBarWriter(
self.dest,
self.trading_calendar,
TEST_CALENDAR_START,
TEST_CALENDAR_STOP,
US_EQUITIES_MINUTES_PER_DAY,
ohlc_ratios_per_sid={sid: 25},
)
writer_with_ratios.write_sid(sid, data)
reader = BcolzMinuteBarReader(self.dest)
open_price = reader.get_value(sid, minute, "open")
self.assertEquals(10.0, open_price)
high_price = reader.get_value(sid, minute, "high")
self.assertEquals(20.0, high_price)
low_price = reader.get_value(sid, minute, "low")
self.assertEquals(30.0, low_price)
close_price = reader.get_value(sid, minute, "close")
self.assertEquals(40.0, close_price)
volume_price = reader.get_value(sid, minute, "volume")
self.assertEquals(50.0, volume_price)
def ingest(
environ: Mapping,
asset_db_writer: AssetDBWriter,
minute_bar_writer: BcolzMinuteBarWriter,
daily_bar_writer: BcolzDailyBarWriter,
adjustment_writer: SQLiteAdjustmentWriter,
calendar: TradingCalendar,
start_session: pd.Timestamp,
end_session: pd.Timestamp,
cache: dataframe_cache,
show_progress: bool,
output_dir: Text,
) -> NoReturn:
sid_map = list(zip(range(len(symbols)), symbols))
asset_db_writer.write(
futures=_get_metadata(sid_map),
exchanges=pd.DataFrame(data=[['bitmex', 'UTC']],
columns=['exchange', 'timezone']),
)
minute_bar_writer.write(
_get_minute_bars(sid_map, start_session, end_session, cache),
show_progress=show_progress,
)
daily_bar_writer.write(
_get_daily_bars(sid_map, start_session, end_session, cache),
show_progress=show_progress,
)
def init_instance_fixtures(self):
super(BcolzMinuteBarTestCase, self).init_instance_fixtures()
self.dest = self.instance_tmpdir.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START,
self.dest,
self.market_opens,
self.market_closes,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def setUp(self):
self.dir_ = TempDirectory()
self.dir_.create()
self.dest = self.dir_.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START,
self.dest,
self.market_opens,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def init_instance_fixtures(self):
super(BcolzMinuteBarTestCase, self).init_instance_fixtures()
self.dest = self.instance_tmpdir.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
self.dest,
self.trading_calendar,
TEST_CALENDAR_START,
TEST_CALENDAR_STOP,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def write_bcolz_minute_data(env, days, path, df_dict):
market_opens = env.open_and_closes.market_open.loc[days]
market_closes = env.open_and_closes.market_close.loc[days]
writer = BcolzMinuteBarWriter(
days[0],
path,
market_opens,
market_closes,
US_EQUITIES_MINUTES_PER_DAY
)
for sid, df in iteritems(df_dict):
writer.write(sid, df)
def test_truncate_between_data_points(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + Timedelta(days=1),
end=self.test_calendar_start + Timedelta(days=3)
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write_sid(sid, data)
# Open a new writer to cover `open` method, also truncating only
# applies to an existing directory.
writer = BcolzMinuteBarWriter.open(self.dest)
# Truncate to first day with data.
writer.truncate(days[0])
# Refresh the reader since truncate update the metadata.
self.reader = BcolzMinuteBarReader(self.dest)
self.assertEqual(self.writer.last_date_in_output_for_sid(sid), days[0])
cal = self.trading_calendar
_, last_close = cal.open_and_close_for_session(days[0])
self.assertEqual(self.reader.last_available_dt, last_close)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_truncate_between_data_points(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write_sid(sid, data)
# Open a new writer to cover `open` method, also truncating only
# applies to an existing directory.
writer = BcolzMinuteBarWriter.open(self.dest)
# Truncate to first day with data.
writer.truncate(days[0])
# Refresh the reader since truncate update the metadata.
self.reader = BcolzMinuteBarReader(self.dest)
self.assertEqual(self.writer.last_date_in_output_for_sid(sid), days[0])
cal = self.trading_calendar
_, last_close = cal.open_and_close_for_session(days[0])
self.assertEqual(self.reader.last_available_dt, last_close)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def write_bcolz_minute_data(trading_calendar, days, path, data):
BcolzMinuteBarWriter(
path,
trading_calendar,
days[0],
days[-1],
US_EQUITIES_MINUTES_PER_DAY
).write(data)
def init_instance_fixtures(self):
super(BcolzMinuteBarTestCase, self).init_instance_fixtures()
self.dest = self.instance_tmpdir.getpath("minute_bars")
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
self.dest, self.trading_calendar, TEST_CALENDAR_START, TEST_CALENDAR_STOP, US_EQUITIES_MINUTES_PER_DAY
)
self.reader = BcolzMinuteBarReader(self.dest)
def setUp(self):
self.dir_ = TempDirectory()
self.dir_.create()
self.dest = self.dir_.getpath("minute_bars")
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START, self.dest, self.market_opens, US_EQUITIES_MINUTES_PER_DAY
)
self.reader = BcolzMinuteBarReader(self.dest)
def write_bcolz_minute_data(env, days, path, data):
market_opens = env.open_and_closes.market_open.loc[days]
market_closes = env.open_and_closes.market_close.loc[days]
BcolzMinuteBarWriter(
days[0],
path,
market_opens,
market_closes,
US_EQUITIES_MINUTES_PER_DAY
).write(data)
def write_bcolz_minute_data(trading_calendar, days, path, data):
market_opens = trading_calendar.schedule.loc[days].market_open
market_closes = trading_calendar.schedule.loc[days].market_close
BcolzMinuteBarWriter(
days[0],
path,
market_opens,
market_closes,
US_EQUITIES_MINUTES_PER_DAY
).write(data)
def test_write_one_ohlcv_with_ratios(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = pd.DataFrame(
data={
"open": [10.0],
"high": [20.0],
"low": [30.0],
"close": [40.0],
"volume": [50.0],
},
index=[minute],
)
# Create a new writer with `ohlc_ratios_per_sid` defined.
writer_with_ratios = BcolzMinuteBarWriter(
self.dest,
self.trading_calendar,
TEST_CALENDAR_START,
TEST_CALENDAR_STOP,
US_EQUITIES_MINUTES_PER_DAY,
ohlc_ratios_per_sid={sid: 25},
)
writer_with_ratios.write_sid(sid, data)
reader = BcolzMinuteBarReader(self.dest)
open_price = reader.get_value(sid, minute, "open")
assert 10.0 == open_price
high_price = reader.get_value(sid, minute, "high")
assert 20.0 == high_price
low_price = reader.get_value(sid, minute, "low")
assert 30.0 == low_price
close_price = reader.get_value(sid, minute, "close")
assert 40.0 == close_price
volume_price = reader.get_value(sid, minute, "volume")
assert 50.0 == volume_price
def test_write_one_ohlcv_with_ratios(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0],
},
index=[minute],
)
# Create a new writer with `ohlc_ratios_per_sid` defined.
writer_with_ratios = BcolzMinuteBarWriter(
self.dest,
self.trading_calendar,
TEST_CALENDAR_START,
TEST_CALENDAR_STOP,
US_EQUITIES_MINUTES_PER_DAY,
ohlc_ratios_per_sid={sid: 25},
)
writer_with_ratios.write_sid(sid, data)
reader = BcolzMinuteBarReader(self.dest)
open_price = reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def init_instance_fixtures(self):
super(BcolzMinuteBarTestCase, self).init_instance_fixtures()
self.dest = self.instance_tmpdir.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START,
self.dest,
self.market_opens,
self.market_closes,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def test_append_on_new_day(self):
sid = 1
ohlcv = {
'open': [2.0],
'high': [3.0],
'low': [1.0],
'close': [2.0],
'volume': [10.0]
}
dt = self.market_opens[TEST_CALENDAR_STOP]
data = DataFrame(
data=ohlcv,
index=[dt])
self.writer.write_sid(sid, data)
# Open a new writer to cover `open` method, also a common usage
# of appending new days will be writing to an existing directory.
cday = self.trading_calendar.schedule.index.freq
new_end_session = TEST_CALENDAR_STOP + cday
writer = BcolzMinuteBarWriter.open(self.dest, new_end_session)
next_day_minute = dt + cday
new_data = DataFrame(
data=ohlcv,
index=[next_day_minute])
writer.write_sid(sid, new_data)
# Get a new reader to test updated calendar.
reader = BcolzMinuteBarReader(self.dest)
second_minute = dt + Timedelta(minutes=1)
# The second minute should have been padded with zeros
for col in ('open', 'high', 'low', 'close'):
assert_almost_equal(
nan, reader.get_value(sid, second_minute, col)
)
self.assertEqual(
0, reader.get_value(sid, second_minute, 'volume')
)
# The next day minute should have data.
for col in ('open', 'high', 'low', 'close', 'volume'):
assert_almost_equal(
ohlcv[col], reader.get_value(sid, next_day_minute, col)
)
def test_append_on_new_day(self):
sid = 1
ohlcv = {
'open': [2.0],
'high': [3.0],
'low': [1.0],
'close': [2.0],
'volume': [10.0]
}
dt = self.market_opens[TEST_CALENDAR_STOP]
data = DataFrame(
data=ohlcv,
index=[dt])
self.writer.write_sid(sid, data)
# Open a new writer to cover `open` method, also a common usage
# of appending new days will be writing to an existing directory.
cday = self.trading_calendar.schedule.index.freq
new_end_session = TEST_CALENDAR_STOP + cday
writer = BcolzMinuteBarWriter.open(self.dest, new_end_session)
next_day_minute = dt + cday
new_data = DataFrame(
data=ohlcv,
index=[next_day_minute])
writer.write_sid(sid, new_data)
# Get a new reader to test updated calendar.
reader = BcolzMinuteBarReader(self.dest)
second_minute = dt + Timedelta(minutes=1)
# The second minute should have been padded with zeros
for col in ('open', 'high', 'low', 'close'):
assert_almost_equal(
nan, reader.get_value(sid, second_minute, col)
)
self.assertEqual(
0, reader.get_value(sid, second_minute, 'volume')
)
# The next day minute should have data.
for col in ('open', 'high', 'low', 'close', 'volume'):
assert_almost_equal(
ohlcv[col], reader.get_value(sid, next_day_minute, col)
)
def setUpClass(cls):
cls.env = TradingEnvironment()
cls.tempdir = TempDirectory()
cls.trading_days = cls.env.days_in_range(start=pd.Timestamp(
"2016-01-05", tz='UTC'),
end=pd.Timestamp("2016-01-28",
tz='UTC'))
equities_data = {}
for sid in [1, 2, 3]:
equities_data[sid] = {
"start_date": cls.trading_days[0],
"end_date": cls.env.next_trading_day(cls.trading_days[-1]),
"symbol": "ASSET{0}".format(sid),
}
cls.env.write_data(equities_data=equities_data)
cls.asset1 = cls.env.asset_finder.retrieve_asset(1)
cls.asset2 = cls.env.asset_finder.retrieve_asset(2)
cls.asset3 = cls.env.asset_finder.retrieve_asset(3)
market_opens = cls.env.open_and_closes.market_open.loc[
cls.trading_days]
market_closes = cls.env.open_and_closes.market_close.loc[
cls.trading_days]
minute_writer = BcolzMinuteBarWriter(cls.trading_days[0],
cls.tempdir.path, market_opens,
market_closes,
US_EQUITIES_MINUTES_PER_DAY)
for sid in [1, 2, 3]:
write_minute_data_for_asset(cls.env, minute_writer,
cls.trading_days[0],
cls.trading_days[-1], sid)
cls.adj_reader = cls.create_adjustments_reader()
cls.sim_params = SimulationParameters(period_start=cls.trading_days[0],
period_end=cls.trading_days[-1],
data_frequency="minute",
env=cls.env)
def build_minute_data(cls):
market_opens = cls.env.open_and_closes.market_open.loc[cls.days]
market_closes = cls.env.open_and_closes.market_close.loc[cls.days]
writer = BcolzMinuteBarWriter(cls.days[0], cls.tempdir.path,
market_opens, market_closes,
US_EQUITIES_MINUTES_PER_DAY)
for sid in [cls.ASSET1.sid, cls.SPLIT_ASSET.sid]:
write_minute_data_for_asset(cls.env, writer, cls.days[0],
cls.days[-1], sid)
for sid in [cls.ASSET2.sid, cls.ILLIQUID_SPLIT_ASSET.sid]:
write_minute_data_for_asset(cls.env, writer, cls.days[0],
cls.days[-1], sid, 10)
write_minute_data_for_asset(cls.env, writer, cls.days[0], cls.days[-1],
cls.HILARIOUSLY_ILLIQUID_ASSET.sid, 50)
return BcolzMinuteBarReader(cls.tempdir.path)
class BcolzMinuteBarTestCase(TestCase):
@classmethod
def setUpClass(cls):
cls.env = TradingEnvironment()
all_market_opens = cls.env.open_and_closes.market_open
indexer = all_market_opens.index.slice_indexer(
start=TEST_CALENDAR_START, end=TEST_CALENDAR_STOP)
cls.market_opens = all_market_opens[indexer]
cls.test_calendar_start = cls.market_opens.index[0]
cls.test_calendar_stop = cls.market_opens.index[-1]
def setUp(self):
self.dir_ = TempDirectory()
self.dir_.create()
self.dest = self.dir_.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START,
self.dest,
self.market_opens,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def tearDown(self):
self.dir_.cleanup()
def test_write_one_ohlcv(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_two_bars(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
data = DataFrame(data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=[minute_0, minute_1])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_write_on_second_day(self):
second_day = self.test_calendar_start + 1
minute = self.market_opens[second_day]
sid = 1
data = DataFrame(data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_empty(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(data={
'open': [0],
'high': [0],
'low': [0],
'close': [0],
'volume': [0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
assert_almost_equal(nan, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
assert_almost_equal(nan, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
assert_almost_equal(nan, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
assert_almost_equal(nan, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
assert_almost_equal(0, volume_price)
def test_write_on_multiple_days(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write(sid, data)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
minute = minutes[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(51.0, volume_price)
def test_no_overwrite(self):
minute = self.market_opens[TEST_CALENDAR_START]
sid = 1
data = DataFrame(data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
with self.assertRaises(BcolzMinuteOverlappingData):
self.writer.write(sid, data)
def ingest(name,
environ=os.environ,
timestamp=None,
assets_versions=(),
show_progress=False):
"""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.
"""
try:
bundle = bundles[name]
except KeyError:
raise UnknownBundle(name)
calendar = get_calendar(bundle.calendar_name)
start_session = bundle.start_session
end_session = bundle.end_session
if start_session is None or start_session < calendar.first_session:
start_session = calendar.first_session
if end_session is None or end_session > calendar.last_session:
end_session = calendar.last_session
if timestamp is None:
timestamp = pd.Timestamp.utcnow()
timestamp = timestamp.tz_convert('utc').tz_localize(None)
timestr = to_bundle_ingest_dirname(timestamp)
cachepath = cache_path(name, environ=environ)
pth.ensure_directory(pth.data_path([name, timestr], environ=environ))
pth.ensure_directory(cachepath)
with dataframe_cache(cachepath, clean_on_failure=False) as cache, \
ExitStack() as stack:
# we use `cleanup_on_failure=False` so that we don't purge the
# cache directory if the load fails in the middle
if bundle.create_writers:
wd = stack.enter_context(
working_dir(pth.data_path([], environ=environ)))
daily_bars_path = wd.ensure_dir(
*daily_equity_relative(name, timestr))
daily_bar_writer = CNBcolzDailyBarWriter(
daily_bars_path,
calendar,
start_session,
end_session,
)
# Do an empty write to ensure that the daily ctables exist
# when we create the SQLiteAdjustmentWriter below. The
# SQLiteAdjustmentWriter needs to open the daily ctables so
# that it can compute the adjustment ratios for the dividends.
daily_bar_writer.write(())
minute_bar_writer = BcolzMinuteBarWriter(
wd.ensure_dir(*minute_equity_relative(name, timestr)),
calendar,
start_session,
end_session,
minutes_per_day=bundle.minutes_per_day,
)
assets_db_path = wd.getpath(*asset_db_relative(name, timestr))
asset_db_writer = AssetDBWriter(assets_db_path)
adjustment_db_writer = stack.enter_context(
SQLiteAdjustmentWriter(
wd.getpath(*adjustment_db_relative(name, timestr)),
CNBcolzDailyBarReader(daily_bars_path),
overwrite=True,
))
fundamentals_db_writer = stack.enter_context(
SQLiteFundamentalsWriter(
wd.getpath(*fundamentals_db_relative(
name, timestr, environ=environ)),
overwrite=True,
))
else:
daily_bar_writer = None
minute_bar_writer = None
asset_db_writer = None
adjustment_db_writer = None
if assets_versions:
raise ValueError('Need to ingest a bundle that creates '
'writers in order to downgrade the assets'
' db.')
log.info("Ingesting {}.", name)
bundle.ingest(
environ,
asset_db_writer,
minute_bar_writer,
daily_bar_writer,
adjustment_db_writer,
fundamentals_db_writer,
calendar,
start_session,
end_session,
cache,
show_progress,
pth.data_path([name, timestr], environ=environ),
)
for version in sorted(set(assets_versions), reverse=True):
version_path = wd.getpath(*asset_db_relative(
name,
timestr,
db_version=version,
))
with working_file(version_path) as wf:
shutil.copy2(assets_db_path, wf.path)
downgrade(wf.path, version)
class BcolzMinuteBarTestCase(WithTradingCalendars,
WithInstanceTmpDir,
ZiplineTestCase):
@classmethod
def init_class_fixtures(cls):
super(BcolzMinuteBarTestCase, cls).init_class_fixtures()
cal = cls.trading_calendar.schedule.loc[
TEST_CALENDAR_START:TEST_CALENDAR_STOP
]
cls.market_opens = cal.market_open
cls.market_closes = cal.market_close
cls.test_calendar_start = cls.market_opens.index[0]
cls.test_calendar_stop = cls.market_opens.index[-1]
def init_instance_fixtures(self):
super(BcolzMinuteBarTestCase, self).init_instance_fixtures()
self.dest = self.instance_tmpdir.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
self.dest,
self.trading_calendar,
TEST_CALENDAR_START,
TEST_CALENDAR_STOP,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def test_version(self):
metadata = self.reader._get_metadata()
self.assertEquals(
metadata.version,
BcolzMinuteBarMetadata.FORMAT_VERSION,
)
def test_write_one_ohlcv(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_two_bars(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=[minute_0, minute_1])
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_write_on_second_day(self):
second_day = self.test_calendar_start + 1
minute = self.market_opens[second_day]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_empty(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
'open': [0],
'high': [0],
'low': [0],
'close': [0],
'volume': [0]
},
index=[minute])
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
assert_almost_equal(nan, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
assert_almost_equal(nan, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
assert_almost_equal(nan, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
assert_almost_equal(nan, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
assert_almost_equal(0, volume_price)
def test_write_on_multiple_days(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write_sid(sid, data)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
minute = minutes[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(51.0, volume_price)
def test_no_overwrite(self):
minute = self.market_opens[TEST_CALENDAR_START]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write_sid(sid, data)
with self.assertRaises(BcolzMinuteOverlappingData):
self.writer.write_sid(sid, data)
def test_append_to_same_day(self):
"""
Test writing data with the same date as existing data in our file.
"""
sid = 1
first_minute = self.market_opens[TEST_CALENDAR_START]
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[first_minute])
self.writer.write_sid(sid, data)
# Write data in the same day as the previous minute
second_minute = first_minute + Timedelta(minutes=1)
new_data = DataFrame(
data={
'open': [5.0],
'high': [10.0],
'low': [3.0],
'close': [7.0],
'volume': [10.0]
},
index=[second_minute])
self.writer.write_sid(sid, new_data)
open_price = self.reader.get_value(sid, second_minute, 'open')
self.assertEquals(5.0, open_price)
high_price = self.reader.get_value(sid, second_minute, 'high')
self.assertEquals(10.0, high_price)
low_price = self.reader.get_value(sid, second_minute, 'low')
self.assertEquals(3.0, low_price)
close_price = self.reader.get_value(sid, second_minute, 'close')
self.assertEquals(7.0, close_price)
volume_price = self.reader.get_value(sid, second_minute, 'volume')
self.assertEquals(10.0, volume_price)
def test_append_on_new_day(self):
sid = 1
ohlcv = {
'open': [2.0],
'high': [3.0],
'low': [1.0],
'close': [2.0],
'volume': [10.0]
}
first_minute = self.market_opens[TEST_CALENDAR_START]
data = DataFrame(
data=ohlcv,
index=[first_minute])
self.writer.write_sid(sid, data)
next_day_minute = first_minute + Timedelta(days=1)
new_data = DataFrame(
data=ohlcv,
index=[next_day_minute])
self.writer.write_sid(sid, new_data)
second_minute = first_minute + Timedelta(minutes=1)
# The second minute should have been padded with zeros
for col in ('open', 'high', 'low', 'close'):
assert_almost_equal(
nan, self.reader.get_value(sid, second_minute, col)
)
self.assertEqual(
0, self.reader.get_value(sid, second_minute, 'volume')
)
# The first day should contain US_EQUITIES_MINUTES_PER_DAY rows.
# The second day should contain a single row.
self.assertEqual(
len(self.writer._ensure_ctable(sid)),
US_EQUITIES_MINUTES_PER_DAY + 1,
)
def test_write_multiple_sids(self):
"""
Test writing multiple sids.
Tests both that the data is written to the correct sid, as well as
ensuring that the logic for creating the subdirectory path to each sid
does not cause issues from attempts to recreate existing paths.
(Calling out this coverage, because an assertion of that logic does not
show up in the test itself, but is exercised by the act of attempting
to write two consecutive sids, which would be written to the same
containing directory, `00/00/000001.bcolz` and `00/00/000002.bcolz)
Before applying a check to make sure the path writing did not
re-attempt directory creation an OSError like the following would
occur:
```
OSError: [Errno 17] File exists: '/tmp/tmpR7yzzT/minute_bars/00/00'
```
"""
minute = self.market_opens[TEST_CALENDAR_START]
sids = [1, 2]
data = DataFrame(
data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write_sid(sids[0], data)
data = DataFrame(
data={
'open': [25.0],
'high': [27.0],
'low': [21.0],
'close': [25.0],
'volume': [200.0]
},
index=[minute])
self.writer.write_sid(sids[1], data)
sid = sids[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
sid = sids[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(25.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(27.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(21.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(25.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(200.0, volume_price)
def test_pad_data(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
day = TEST_CALENDAR_START + freq
minute = self.market_opens[day]
data = DataFrame(
data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
# Check that if we then pad the rest of this day, we end up with
# 2 days worth of minutes.
self.writer.pad(sid, day)
self.assertEqual(
len(self.writer._ensure_ctable(sid)),
self.writer._minutes_per_day * 2,
)
def test_nans(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
minutes = date_range(minute, periods=9, freq='min')
data = DataFrame(
data={
'open': full(9, nan),
'high': full(9, nan),
'low': full(9, nan),
'close': full(9, nan),
'volume': full(9, 0.0),
},
index=[minutes])
self.writer.write_sid(sid, data)
fields = ['open', 'high', 'low', 'close', 'volume']
ohlcv_window = list(map(transpose, self.reader.load_raw_arrays(
fields, minutes[0], minutes[-1], [sid],
)))
for i, field in enumerate(fields):
if field != 'volume':
assert_array_equal(full(9, nan), ohlcv_window[i][0])
else:
assert_array_equal(zeros(9), ohlcv_window[i][0])
def test_differing_nans(self):
"""
Also test nans of differing values/construction.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
minutes = date_range(minute, periods=9, freq='min')
data = DataFrame(
data={
'open': ((0b11111111111 << 52) + arange(1, 10, dtype=int64)).
view(float64),
'high': ((0b11111111111 << 52) + arange(11, 20, dtype=int64)).
view(float64),
'low': ((0b11111111111 << 52) + arange(21, 30, dtype=int64)).
view(float64),
'close': ((0b11111111111 << 52) + arange(31, 40, dtype=int64)).
view(float64),
'volume': full(9, 0.0),
},
index=[minutes])
self.writer.write_sid(sid, data)
fields = ['open', 'high', 'low', 'close', 'volume']
ohlcv_window = list(map(transpose, self.reader.load_raw_arrays(
fields, minutes[0], minutes[-1], [sid],
)))
for i, field in enumerate(fields):
if field != 'volume':
assert_array_equal(full(9, nan), ohlcv_window[i][0])
else:
assert_array_equal(zeros(9), ohlcv_window[i][0])
def test_write_cols(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
cols = {
'open': array([10.0, 11.0]),
'high': array([20.0, 21.0]),
'low': array([30.0, 31.0]),
'close': array([40.0, 41.0]),
'volume': array([50.0, 51.0])
}
dts = array([minute_0, minute_1], dtype='datetime64[s]')
self.writer.write_cols(sid, dts, cols)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_write_cols_mismatch_length(self):
dts = date_range(self.market_opens[self.test_calendar_start],
periods=2, freq='min').asi8.astype('datetime64[s]')
sid = 1
cols = {
'open': array([10.0, 11.0, 12.0]),
'high': array([20.0, 21.0]),
'low': array([30.0, 31.0, 33.0, 34.0]),
'close': array([40.0, 41.0]),
'volume': array([50.0, 51.0, 52.0])
}
with self.assertRaises(BcolzMinuteWriterColumnMismatch):
self.writer.write_cols(sid, dts, cols)
def test_unadjusted_minutes(self):
"""
Test unadjusted minutes.
"""
start_minute = self.market_opens[TEST_CALENDAR_START]
minutes = [start_minute,
start_minute + Timedelta('1 min'),
start_minute + Timedelta('2 min')]
sids = [1, 2]
data_1 = DataFrame(
data={
'open': [15.0, nan, 15.1],
'high': [17.0, nan, 17.1],
'low': [11.0, nan, 11.1],
'close': [14.0, nan, 14.1],
'volume': [1000, 0, 1001]
},
index=minutes)
self.writer.write_sid(sids[0], data_1)
data_2 = DataFrame(
data={
'open': [25.0, nan, 25.1],
'high': [27.0, nan, 27.1],
'low': [21.0, nan, 21.1],
'close': [24.0, nan, 24.1],
'volume': [2000, 0, 2001]
},
index=minutes)
self.writer.write_sid(sids[1], data_2)
reader = BcolzMinuteBarReader(self.dest)
columns = ['open', 'high', 'low', 'close', 'volume']
sids = [sids[0], sids[1]]
arrays = list(map(transpose, reader.load_raw_arrays(
columns, minutes[0], minutes[-1], sids,
)))
data = {sids[0]: data_1, sids[1]: data_2}
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
assert_almost_equal(data[sid][col], arrays[i][j])
def test_unadjusted_minutes_early_close(self):
"""
Test unadjusted minute window, ensuring that early closes are filtered
out.
"""
day_before_thanksgiving = Timestamp('2015-11-25', tz='UTC')
xmas_eve = Timestamp('2015-12-24', tz='UTC')
market_day_after_xmas = Timestamp('2015-12-28', tz='UTC')
minutes = [self.market_closes[day_before_thanksgiving] -
Timedelta('2 min'),
self.market_closes[xmas_eve] - Timedelta('1 min'),
self.market_opens[market_day_after_xmas] +
Timedelta('1 min')]
sids = [1, 2]
data_1 = DataFrame(
data={
'open': [
15.0, 15.1, 15.2],
'high': [17.0, 17.1, 17.2],
'low': [11.0, 11.1, 11.3],
'close': [14.0, 14.1, 14.2],
'volume': [1000, 1001, 1002],
},
index=minutes)
self.writer.write_sid(sids[0], data_1)
data_2 = DataFrame(
data={
'open': [25.0, 25.1, 25.2],
'high': [27.0, 27.1, 27.2],
'low': [21.0, 21.1, 21.2],
'close': [24.0, 24.1, 24.2],
'volume': [2000, 2001, 2002],
},
index=minutes)
self.writer.write_sid(sids[1], data_2)
reader = BcolzMinuteBarReader(self.dest)
columns = ['open', 'high', 'low', 'close', 'volume']
sids = [sids[0], sids[1]]
arrays = list(map(transpose, reader.load_raw_arrays(
columns, minutes[0], minutes[-1], sids,
)))
data = {sids[0]: data_1, sids[1]: data_2}
start_minute_loc = \
self.trading_calendar.all_minutes.get_loc(minutes[0])
minute_locs = [
self.trading_calendar.all_minutes.get_loc(minute)
- start_minute_loc
for minute in minutes
]
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
assert_almost_equal(data[sid].loc[minutes, col],
arrays[i][j][minute_locs])
def test_adjust_non_trading_minutes(self):
start_day = Timestamp('2015-06-01', tz='UTC')
end_day = Timestamp('2015-06-02', tz='UTC')
sid = 1
cols = {
'open': arange(1, 781),
'high': arange(1, 781),
'low': arange(1, 781),
'close': arange(1, 781),
'volume': arange(1, 781)
}
dts = array(self.trading_calendar.minutes_for_sessions_in_range(
self.trading_calendar.minute_to_session_label(start_day),
self.trading_calendar.minute_to_session_label(end_day)
))
self.writer.write_cols(sid, dts, cols)
self.assertEqual(
self.reader.get_value(
sid,
Timestamp('2015-06-01 20:00:00', tz='UTC'),
'open'),
390)
self.assertEqual(
self.reader.get_value(
sid,
Timestamp('2015-06-02 20:00:00', tz='UTC'),
'open'),
780)
with self.assertRaises(NoDataOnDate):
self.reader.get_value(
sid,
Timestamp('2015-06-02', tz='UTC'),
'open'
)
with self.assertRaises(NoDataOnDate):
self.reader.get_value(
sid,
Timestamp('2015-06-02 20:01:00', tz='UTC'),
'open'
)
def test_adjust_non_trading_minutes_half_days(self):
# half day
start_day = Timestamp('2015-11-27', tz='UTC')
end_day = Timestamp('2015-11-30', tz='UTC')
sid = 1
cols = {
'open': arange(1, 601),
'high': arange(1, 601),
'low': arange(1, 601),
'close': arange(1, 601),
'volume': arange(1, 601)
}
dts = array(
self.trading_calendar.minutes_for_sessions_in_range(
self.trading_calendar.minute_to_session_label(start_day),
self.trading_calendar.minute_to_session_label(end_day)
)
)
self.writer.write_cols(sid, dts, cols)
self.assertEqual(
self.reader.get_value(
sid,
Timestamp('2015-11-27 18:00:00', tz='UTC'),
'open'),
210)
self.assertEqual(
self.reader.get_value(
sid,
Timestamp('2015-11-30 21:00:00', tz='UTC'),
'open'),
600)
self.assertEqual(
self.reader.get_value(
sid,
Timestamp('2015-11-27 18:01:00', tz='UTC'),
'open'),
210)
with self.assertRaises(NoDataOnDate):
self.reader.get_value(
sid,
Timestamp('2015-11-30', tz='UTC'),
'open'
)
with self.assertRaises(NoDataOnDate):
self.reader.get_value(
sid,
Timestamp('2015-11-30 21:01:00', tz='UTC'),
'open'
)
def test_set_sid_attrs(self):
"""Confirm that we can set the attributes of a sid's file correctly.
"""
sid = 1
start_day = Timestamp('2015-11-27', tz='UTC')
end_day = Timestamp('2015-06-02', tz='UTC')
attrs = {
'start_day': start_day.value / int(1e9),
'end_day': end_day.value / int(1e9),
'factor': 100,
}
# Write the attributes
self.writer.set_sid_attrs(sid, **attrs)
# Read the attributes
for k, v in attrs.items():
self.assertEqual(self.reader.get_sid_attr(sid, k), v)
def test_truncate_between_data_points(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write_sid(sid, data)
# Truncate to first day with data.
self.writer.truncate(days[0])
# Refresh the reader since truncate update the metadata.
self.reader = BcolzMinuteBarReader(self.dest)
self.assertEqual(self.writer.last_date_in_output_for_sid(sid), days[0])
cal = self.trading_calendar
_, last_close = cal.open_and_close_for_session(days[0])
self.assertEqual(self.reader.last_available_dt, last_close)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_truncate_all_data_points(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write_sid(sid, data)
# Truncate to first day in the calendar, a day before the first
# day with minute data.
self.writer.truncate(self.test_calendar_start)
# Refresh the reader since truncate update the metadata.
self.reader = BcolzMinuteBarReader(self.dest)
self.assertEqual(
self.writer.last_date_in_output_for_sid(sid),
self.test_calendar_start,
)
cal = self.trading_calendar
_, last_close = cal.open_and_close_for_session(
self.test_calendar_start)
self.assertEqual(self.reader.last_available_dt, last_close)
def test_unadjusted_minutes_market_breaks(self):
"""
Test unadjusted minute window, ensuring that market breaks are filtered
out.
"""
MINUTES_PER_DAY = 360
trading_calendar = get_calendar("XTKS")
writer = BcolzMinuteBarWriter(
self.dest,
trading_calendar,
TEST_CALENDAR_START,
TEST_CALENDAR_STOP,
MINUTES_PER_DAY,
)
sample_date = Timestamp('2015-11-25', tz='UTC')
minutes = [
# before break
trading_calendar.break_starts[sample_date] - Timedelta('1 min'),
# after break
trading_calendar.break_ends[sample_date] + Timedelta('1 min'),
trading_calendar.break_ends[sample_date] + Timedelta('2 min')
]
sids = [1, 2]
data_1 = DataFrame(data={
'open': [15.0, 15.1, 15.2],
'high': [17.0, 17.1, 17.2],
'low': [11.0, 11.1, 11.3],
'close': [14.0, 14.1, 14.2],
'volume': [1000, 1001, 1002],
},
index=minutes)
writer.write_sid(sids[0], data_1)
data_2 = DataFrame(data={
'open': [25.0, 25.1, 25.2],
'high': [27.0, 27.1, 27.2],
'low': [21.0, 21.1, 21.2],
'close': [24.0, 24.1, 24.2],
'volume': [2000, 2001, 2002],
},
index=minutes)
writer.write_sid(sids[1], data_2)
reader = BcolzMinuteBarReader(self.dest)
columns = ['open', 'high', 'low', 'close', 'volume']
sids = [sids[0], sids[1]]
arrays = list(
map(
transpose,
reader.load_raw_arrays(
columns,
minutes[0],
minutes[-1],
sids,
)))
print(arrays)
data = {sids[0]: data_1, sids[1]: data_2}
print('data')
print(data)
start_minute_loc = \
trading_calendar.all_minutes.get_loc(minutes[0])
minute_locs = [
trading_calendar.all_minutes.get_loc(minute) - start_minute_loc
for minute in minutes
]
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
print('sid', sid, 'col', col)
print('i', i, 'j', j)
print(minutes)
print(minute_locs)
expected = data[sid].loc[minutes, col]
actual = arrays[i][j][minute_locs]
assert_almost_equal(expected, actual)
def test_truncate_between_data_points(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + timedelta(days=1),
end=self.test_calendar_start + timedelta(days=3),
)]
minutes = pd.DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = pd.DataFrame(
data={
"open": [10.0, 11.0],
"high": [20.0, 21.0],
"low": [30.0, 31.0],
"close": [40.0, 41.0],
"volume": [50.0, 51.0],
},
index=minutes,
)
self.writer.write_sid(sid, data)
# Open a new writer to cover `open` method, also truncating only
# applies to an existing directory.
writer = BcolzMinuteBarWriter.open(self.dest)
# Truncate to first day with data.
writer.truncate(days[0])
# Refresh the reader since truncate update the metadata.
self.reader = BcolzMinuteBarReader(self.dest)
assert self.writer.last_date_in_output_for_sid(sid) == days[0]
cal = self.trading_calendar
_, last_close = cal.open_and_close_for_session(days[0])
assert self.reader.last_available_dt == last_close
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, "open")
assert 10.0 == open_price
high_price = self.reader.get_value(sid, minute, "high")
assert 20.0 == high_price
low_price = self.reader.get_value(sid, minute, "low")
assert 30.0 == low_price
close_price = self.reader.get_value(sid, minute, "close")
assert 40.0 == close_price
volume_price = self.reader.get_value(sid, minute, "volume")
assert 50.0 == volume_price
class BcolzMinuteBarTestCase(TestCase):
@classmethod
def setUpClass(cls):
cls.env = TradingEnvironment()
all_market_opens = cls.env.open_and_closes.market_open
all_market_closes = cls.env.open_and_closes.market_close
indexer = all_market_opens.index.slice_indexer(
start=TEST_CALENDAR_START,
end=TEST_CALENDAR_STOP
)
cls.market_opens = all_market_opens[indexer]
cls.market_closes = all_market_closes[indexer]
cls.test_calendar_start = cls.market_opens.index[0]
cls.test_calendar_stop = cls.market_opens.index[-1]
def setUp(self):
self.dir_ = TempDirectory()
self.dir_.create()
self.dest = self.dir_.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START,
self.dest,
self.market_opens,
self.market_closes,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def tearDown(self):
self.dir_.cleanup()
def test_write_one_ohlcv(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_two_bars(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=[minute_0, minute_1])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_write_on_second_day(self):
second_day = self.test_calendar_start + 1
minute = self.market_opens[second_day]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_empty(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
'open': [0],
'high': [0],
'low': [0],
'close': [0],
'volume': [0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
assert_almost_equal(nan, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
assert_almost_equal(nan, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
assert_almost_equal(nan, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
assert_almost_equal(nan, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
assert_almost_equal(0, volume_price)
def test_write_on_multiple_days(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write(sid, data)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
minute = minutes[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(51.0, volume_price)
def test_no_overwrite(self):
minute = self.market_opens[TEST_CALENDAR_START]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
with self.assertRaises(BcolzMinuteOverlappingData):
self.writer.write(sid, data)
def test_write_multiple_sids(self):
"""
Test writing multiple sids.
Tests both that the data is written to the correct sid, as well as
ensuring that the logic for creating the subdirectory path to each sid
does not cause issues from attempts to recreate existing paths.
(Calling out this coverage, because an assertion of that logic does not
show up in the test itself, but is exercised by the act of attempting
to write two consecutive sids, which would be written to the same
containing directory, `00/00/000001.bcolz` and `00/00/000002.bcolz)
Before applying a check to make sure the path writing did not
re-attempt directory creation an OSError like the following would
occur:
```
OSError: [Errno 17] File exists: '/tmp/tmpR7yzzT/minute_bars/00/00'
```
"""
minute = self.market_opens[TEST_CALENDAR_START]
sids = [1, 2]
data = DataFrame(
data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write(sids[0], data)
data = DataFrame(
data={
'open': [25.0],
'high': [27.0],
'low': [21.0],
'close': [25.0],
'volume': [200.0]
},
index=[minute])
self.writer.write(sids[1], data)
sid = sids[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
sid = sids[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(25.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(27.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(21.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(25.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(200.0, volume_price)
def test_pad_data(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
data = DataFrame(
data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
def test_nans(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
minutes = date_range(minute, periods=9, freq='min')
data = DataFrame(
data={
'open': full(9, nan),
'high': full(9, nan),
'low': full(9, nan),
'close': full(9, nan),
'volume': full(9, 0),
},
index=[minutes])
self.writer.write(sid, data)
fields = ['open', 'high', 'low', 'close', 'volume']
ohlcv_window = self.reader.unadjusted_window(
fields, minutes[0], minutes[-1], [sid])
for i, field in enumerate(fields):
if field != 'volume':
assert_array_equal(full(9, nan), ohlcv_window[i][0])
else:
assert_array_equal(zeros(9), ohlcv_window[i][0])
def test_differing_nans(self):
"""
Also test nans of differing values/construction.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
minutes = date_range(minute, periods=9, freq='min')
data = DataFrame(
data={
'open': ((0b11111111111 << 52) + arange(1, 10, dtype=int64)).
view(float64),
'high': ((0b11111111111 << 52) + arange(11, 20, dtype=int64)).
view(float64),
'low': ((0b11111111111 << 52) + arange(21, 30, dtype=int64)).
view(float64),
'close': ((0b11111111111 << 52) + arange(31, 40, dtype=int64)).
view(float64),
'volume': full(9, 0),
},
index=[minutes])
self.writer.write(sid, data)
fields = ['open', 'high', 'low', 'close', 'volume']
ohlcv_window = self.reader.unadjusted_window(
fields, minutes[0], minutes[-1], [sid])
for i, field in enumerate(fields):
if field != 'volume':
assert_array_equal(full(9, nan), ohlcv_window[i][0])
else:
assert_array_equal(zeros(9), ohlcv_window[i][0])
def test_write_cols(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
cols = {
'open': array([10.0, 11.0]),
'high': array([20.0, 21.0]),
'low': array([30.0, 31.0]),
'close': array([40.0, 41.0]),
'volume': array([50.0, 51.0])
}
dts = array([minute_0, minute_1], dtype='datetime64[s]')
self.writer.write_cols(sid, dts, cols)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_unadjusted_minutes(self):
"""
Test unadjusted minutes.
"""
start_minute = self.market_opens[TEST_CALENDAR_START]
minutes = [start_minute,
start_minute + Timedelta('1 min'),
start_minute + Timedelta('2 min')]
sids = [1, 2]
data_1 = DataFrame(
data={
'open': [15.0, nan, 15.1],
'high': [17.0, nan, 17.1],
'low': [11.0, nan, 11.1],
'close': [14.0, nan, 14.1],
'volume': [1000, 0, 1001]
},
index=minutes)
self.writer.write(sids[0], data_1)
data_2 = DataFrame(
data={
'open': [25.0, nan, 25.1],
'high': [27.0, nan, 27.1],
'low': [21.0, nan, 21.1],
'close': [24.0, nan, 24.1],
'volume': [2000, 0, 2001]
},
index=minutes)
self.writer.write(sids[1], data_2)
reader = BcolzMinuteBarReader(self.dest)
columns = ['open', 'high', 'low', 'close', 'volume']
sids = [sids[0], sids[1]]
arrays = reader.unadjusted_window(
columns, minutes[0], minutes[-1], sids)
data = {sids[0]: data_1, sids[1]: data_2}
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
assert_almost_equal(data[sid][col], arrays[i][j])
class BcolzMinuteBarTestCase(TestCase):
@classmethod
def setUpClass(cls):
cls.env = TradingEnvironment()
all_market_opens = cls.env.open_and_closes.market_open
indexer = all_market_opens.index.slice_indexer(
start=TEST_CALENDAR_START,
end=TEST_CALENDAR_STOP
)
cls.market_opens = all_market_opens[indexer]
cls.test_calendar_start = cls.market_opens.index[0]
cls.test_calendar_stop = cls.market_opens.index[-1]
def setUp(self):
self.dir_ = TempDirectory()
self.dir_.create()
self.dest = self.dir_.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START,
self.dest,
self.market_opens,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def tearDown(self):
self.dir_.cleanup()
def test_write_one_ohlcv(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_two_bars(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=[minute_0, minute_1])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_write_on_second_day(self):
second_day = self.test_calendar_start + 1
minute = self.market_opens[second_day]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_empty(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
'open': [0],
'high': [0],
'low': [0],
'close': [0],
'volume': [0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
assert_almost_equal(nan, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
assert_almost_equal(nan, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
assert_almost_equal(nan, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
assert_almost_equal(nan, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
assert_almost_equal(0, volume_price)
def test_write_on_multiple_days(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write(sid, data)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
minute = minutes[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(51.0, volume_price)
def test_no_overwrite(self):
minute = self.market_opens[TEST_CALENDAR_START]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
with self.assertRaises(BcolzMinuteOverlappingData):
self.writer.write(sid, data)
def test_write_multiple_sids(self):
"""
Test writing multiple sids.
Tests both that the data is written to the correct sid, as well as
ensuring that the logic for creating the subdirectory path to each sid
does not cause issues from attempts to recreate existing paths.
(Calling out this coverage, because an assertion of that logic does not
show up in the test itself, but is exercised by the act of attempting
to write two consecutive sids, which would be written to the same
containing directory, `00/00/000001.bcolz` and `00/00/000002.bcolz)
Before applying a check to make sure the path writing did not
re-attempt directory creation an OSError like the following would
occur:
```
OSError: [Errno 17] File exists: '/tmp/tmpR7yzzT/minute_bars/00/00'
```
"""
minute = self.market_opens[TEST_CALENDAR_START]
sids = [1, 2]
data = DataFrame(
data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write(sids[0], data)
data = DataFrame(
data={
'open': [25.0],
'high': [27.0],
'low': [21.0],
'close': [25.0],
'volume': [200.0]
},
index=[minute])
self.writer.write(sids[1], data)
sid = sids[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
sid = sids[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(25.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(27.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(21.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(25.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(200.0, volume_price)
def test_pad_data(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
data = DataFrame(
data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
def test_write_cols(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
cols = {
'open': array([10.0, 11.0]),
'high': array([20.0, 21.0]),
'low': array([30.0, 31.0]),
'close': array([40.0, 41.0]),
'volume': array([50.0, 51.0])
}
dts = array([minute_0, minute_1], dtype='datetime64[s]')
self.writer.write_cols(sid, dts, cols)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
class BcolzMinuteBarTestCase(TestCase):
@classmethod
def setUpClass(cls):
cls.env = TradingEnvironment()
all_market_opens = cls.env.open_and_closes.market_open
indexer = all_market_opens.index.slice_indexer(
start=TEST_CALENDAR_START,
end=TEST_CALENDAR_STOP
)
cls.market_opens = all_market_opens[indexer]
cls.test_calendar_start = cls.market_opens.index[0]
cls.test_calendar_stop = cls.market_opens.index[-1]
def setUp(self):
self.dir_ = TempDirectory()
self.dir_.create()
self.dest = self.dir_.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START,
self.dest,
self.market_opens,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def tearDown(self):
self.dir_.cleanup()
def test_write_one_ohlcv(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_two_bars(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=[minute_0, minute_1])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_write_on_second_day(self):
second_day = self.test_calendar_start + 1
minute = self.market_opens[second_day]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_empty(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
'open': [0],
'high': [0],
'low': [0],
'close': [0],
'volume': [0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
assert_almost_equal(nan, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
assert_almost_equal(nan, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
assert_almost_equal(nan, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
assert_almost_equal(nan, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
assert_almost_equal(0, volume_price)
def test_write_on_multiple_days(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write(sid, data)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
minute = minutes[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(51.0, volume_price)
def test_no_overwrite(self):
minute = self.market_opens[TEST_CALENDAR_START]
sid = 1
data = DataFrame(
data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
with self.assertRaises(BcolzMinuteOverlappingData):
self.writer.write(sid, data)
class BcolzMinuteBarTestCase(TestCase):
@classmethod
def setUpClass(cls):
cls.env = TradingEnvironment()
all_market_opens = cls.env.open_and_closes.market_open
indexer = all_market_opens.index.slice_indexer(
start=TEST_CALENDAR_START, end=TEST_CALENDAR_STOP)
cls.market_opens = all_market_opens[indexer]
cls.test_calendar_start = cls.market_opens.index[0]
cls.test_calendar_stop = cls.market_opens.index[-1]
def setUp(self):
self.dir_ = TempDirectory()
self.dir_.create()
self.dest = self.dir_.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START,
self.dest,
self.market_opens,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def tearDown(self):
self.dir_.cleanup()
def test_write_one_ohlcv(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_two_bars(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
data = DataFrame(data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=[minute_0, minute_1])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_write_on_second_day(self):
second_day = self.test_calendar_start + 1
minute = self.market_opens[second_day]
sid = 1
data = DataFrame(data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_empty(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(data={
'open': [0],
'high': [0],
'low': [0],
'close': [0],
'volume': [0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
assert_almost_equal(nan, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
assert_almost_equal(nan, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
assert_almost_equal(nan, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
assert_almost_equal(nan, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
assert_almost_equal(0, volume_price)
def test_write_on_multiple_days(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write(sid, data)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
minute = minutes[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(51.0, volume_price)
def test_no_overwrite(self):
minute = self.market_opens[TEST_CALENDAR_START]
sid = 1
data = DataFrame(data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
with self.assertRaises(BcolzMinuteOverlappingData):
self.writer.write(sid, data)
def test_write_multiple_sids(self):
"""
Test writing multiple sids.
Tests both that the data is written to the correct sid, as well as
ensuring that the logic for creating the subdirectory path to each sid
does not cause issues from attempts to recreate existing paths.
(Calling out this coverage, because an assertion of that logic does not
show up in the test itself, but is exercised by the act of attempting
to write two consecutive sids, which would be written to the same
containing directory, `00/00/000001.bcolz` and `00/00/000002.bcolz)
Before applying a check to make sure the path writing did not
re-attempt directory creation an OSError like the following would
occur:
```
OSError: [Errno 17] File exists: '/tmp/tmpR7yzzT/minute_bars/00/00'
```
"""
minute = self.market_opens[TEST_CALENDAR_START]
sids = [1, 2]
data = DataFrame(data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write(sids[0], data)
data = DataFrame(data={
'open': [25.0],
'high': [27.0],
'low': [21.0],
'close': [25.0],
'volume': [200.0]
},
index=[minute])
self.writer.write(sids[1], data)
sid = sids[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
sid = sids[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(25.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(27.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(21.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(25.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(200.0, volume_price)
def test_pad_data(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
data = DataFrame(data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
def test_write_cols(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
cols = {
'open': array([10.0, 11.0]),
'high': array([20.0, 21.0]),
'low': array([30.0, 31.0]),
'close': array([40.0, 41.0]),
'volume': array([50.0, 51.0])
}
dts = array([minute_0, minute_1], dtype='datetime64[s]')
self.writer.write_cols(sid, dts, cols)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_unadjusted_minutes(self):
"""
Test unadjusted minutes.
"""
start_minute = self.market_opens[TEST_CALENDAR_START]
minutes = [
start_minute, start_minute + Timedelta('1 min'),
start_minute + Timedelta('2 min')
]
sids = [1, 2]
data_1 = DataFrame(data={
'open': [15.0, nan, 15.1],
'high': [17.0, nan, 17.1],
'low': [11.0, nan, 11.1],
'close': [14.0, nan, 14.1],
'volume': [1000, 0, 1001]
},
index=minutes)
self.writer.write(sids[0], data_1)
data_2 = DataFrame(data={
'open': [25.0, nan, 25.1],
'high': [27.0, nan, 27.1],
'low': [21.0, nan, 21.1],
'close': [24.0, nan, 24.1],
'volume': [2000, 0, 2001]
},
index=minutes)
self.writer.write(sids[1], data_2)
reader = BcolzMinuteBarReader(self.dest)
columns = ['open', 'high', 'low', 'close', 'volume']
sids = [sids[0], sids[1]]
arrays = reader.unadjusted_window(columns, minutes[0], minutes[-1],
sids)
data = {sids[0]: data_1, sids[1]: data_2}
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
assert_almost_equal(data[sid][col], arrays[i][j])
def fills2reader(self, tempdir, minutes, fills, orders):
if len(minutes)==0:
return None
for _,fill in fills.items():
fill["open"] = fill["close"]
fill["high"] = fill["close"]
fill["low"] = fill["close"]
# since the below abs affects the original dataframe, storing the sign for later revert
fill["is_neg"] = fill["volume"]<0
# take absolute value, since negatives are split in the factory function to begin with
# and zipline doesnt support negative OHLC volumes (which dont make sense anyway)
fill["volume"] = abs(fill["volume"])
# append empty OHLC dataframes for sid's in orders but not (yet) in fills
# dummy OHLC data with volume=0 so as not to affect orders
empty = {"open":[0], "high":[0], "low":[0], "close":[0], "volume":[0], "dt":[minutes[0]], "is_neg":[False]}
for sid in orders:
if sid not in fills:
fills[sid]=pd.DataFrame(empty).set_index("dt")
d1 = self.trading_calendar.minute_to_session_label(
minutes[0]
)
d2=self.trading_calendar.minute_to_session_label(
minutes[-1]
)
days = self.trading_calendar.sessions_in_range(d1, d2)
#print("minutes",minutes)
#print("days: %s, %s, %s" % (d1, d2, days))
#path = os.path.join(tempdir.path, "testdata.bcolz")
path = tempdir.path
writer = BcolzMinuteBarWriter(
rootdir=path,
calendar=self.trading_calendar,
start_session=days[0],
end_session=days[-1],
minutes_per_day=1440
)
#print("Writer session labels: %s" % (writer._session_labels))
#print('last date for sid 1', writer.last_date_in_output_for_sid(1))
#print('last date for sid 2', writer.last_date_in_output_for_sid(2))
#for f in iteritems(fills): print("fill",f)
writer.write(iteritems(fills))
# now that the data is written, revert the volume sign and drop the extra columns
for _,fill in fills.items():
del fill["open"]
del fill["high"]
del fill["low"]
if any(fill["is_neg"]):
fill.loc[fill["is_neg"],"volume"] = -1 * fill["volume"]
del fill["is_neg"]
#print("temp path: %s" % (path))
reader = BcolzMinuteBarReader(path)
return reader
class BcolzMinuteBarTestCase(WithTradingCalendars, WithAssetFinder,
WithInstanceTmpDir, ZiplineTestCase):
ASSET_FINDER_EQUITY_SIDS = 1, 2
@classmethod
def init_class_fixtures(cls):
super(BcolzMinuteBarTestCase, cls).init_class_fixtures()
cal = cls.trading_calendar.schedule.loc[
TEST_CALENDAR_START:TEST_CALENDAR_STOP]
cls.market_opens = cal.market_open.dt.tz_localize("UTC")
cls.market_closes = cal.market_close.dt.tz_localize("UTC")
cls.test_calendar_start = cls.market_opens.index[0]
cls.test_calendar_stop = cls.market_opens.index[-1]
def init_instance_fixtures(self):
super(BcolzMinuteBarTestCase, self).init_instance_fixtures()
self.dest = self.instance_tmpdir.getpath("minute_bars")
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
self.dest,
self.trading_calendar,
TEST_CALENDAR_START,
TEST_CALENDAR_STOP,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def test_version(self):
metadata = self.reader._get_metadata()
self.assertEquals(
metadata.version,
BcolzMinuteBarMetadata.FORMAT_VERSION,
)
def test_no_minute_bars_for_sid(self):
minute = self.market_opens[self.test_calendar_start]
with self.assertRaises(NoDataForSid):
self.reader.get_value(1337, minute, "close")
def test_write_one_ohlcv(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
"open": [10.0],
"high": [20.0],
"low": [30.0],
"close": [40.0],
"volume": [50.0],
},
index=[minute],
)
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, "open")
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, "high")
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, "low")
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, "close")
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
self.assertEquals(50.0, volume_price)
def test_precision_after_scaling(self):
"""For numbers that don't have an exact float representation,
assert that scaling the value does not cause a loss in precision.
"""
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
"open": [130.23],
"high": [130.23],
"low": [130.23],
"close": [130.23],
"volume": [1000],
},
index=[minute],
)
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, "open")
self.assertEquals(130.23, open_price)
high_price = self.reader.get_value(sid, minute, "high")
self.assertEquals(130.23, high_price)
low_price = self.reader.get_value(sid, minute, "low")
self.assertEquals(130.23, low_price)
close_price = self.reader.get_value(sid, minute, "close")
self.assertEquals(130.23, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
self.assertEquals(1000, volume_price)
def test_write_one_ohlcv_with_ratios(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
"open": [10.0],
"high": [20.0],
"low": [30.0],
"close": [40.0],
"volume": [50.0],
},
index=[minute],
)
# Create a new writer with `ohlc_ratios_per_sid` defined.
writer_with_ratios = BcolzMinuteBarWriter(
self.dest,
self.trading_calendar,
TEST_CALENDAR_START,
TEST_CALENDAR_STOP,
US_EQUITIES_MINUTES_PER_DAY,
ohlc_ratios_per_sid={sid: 25},
)
writer_with_ratios.write_sid(sid, data)
reader = BcolzMinuteBarReader(self.dest)
open_price = reader.get_value(sid, minute, "open")
self.assertEquals(10.0, open_price)
high_price = reader.get_value(sid, minute, "high")
self.assertEquals(20.0, high_price)
low_price = reader.get_value(sid, minute, "low")
self.assertEquals(30.0, low_price)
close_price = reader.get_value(sid, minute, "close")
self.assertEquals(40.0, close_price)
volume_price = reader.get_value(sid, minute, "volume")
self.assertEquals(50.0, volume_price)
def test_write_two_bars(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
data = DataFrame(
data={
"open": [10.0, 11.0],
"high": [20.0, 21.0],
"low": [30.0, 31.0],
"close": [40.0, 41.0],
"volume": [50.0, 51.0],
},
index=[minute_0, minute_1],
)
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute_0, "open")
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, "high")
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, "low")
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, "close")
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, "volume")
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, "open")
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, "high")
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, "low")
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, "close")
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, "volume")
self.assertEquals(51.0, volume_price)
def test_write_on_second_day(self):
second_day = self.test_calendar_start + timedelta(days=1)
minute = self.market_opens[second_day]
sid = 1
data = DataFrame(
data={
"open": [10.0],
"high": [20.0],
"low": [30.0],
"close": [40.0],
"volume": [50.0],
},
index=[minute],
)
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, "open")
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, "high")
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, "low")
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, "close")
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
self.assertEquals(50.0, volume_price)
def test_write_empty(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(
data={
"open": [0],
"high": [0],
"low": [0],
"close": [0],
"volume": [0]
},
index=[minute],
)
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, "open")
assert_almost_equal(nan, open_price)
high_price = self.reader.get_value(sid, minute, "high")
assert_almost_equal(nan, high_price)
low_price = self.reader.get_value(sid, minute, "low")
assert_almost_equal(nan, low_price)
close_price = self.reader.get_value(sid, minute, "close")
assert_almost_equal(nan, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
assert_almost_equal(0, volume_price)
def test_write_on_multiple_days(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + timedelta(days=1),
end=self.test_calendar_start + timedelta(days=3),
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
"open": [10.0, 11.0],
"high": [20.0, 21.0],
"low": [30.0, 31.0],
"close": [40.0, 41.0],
"volume": [50.0, 51.0],
},
index=minutes,
)
self.writer.write_sid(sid, data)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, "open")
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, "high")
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, "low")
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, "close")
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
self.assertEquals(50.0, volume_price)
minute = minutes[1]
open_price = self.reader.get_value(sid, minute, "open")
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute, "high")
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute, "low")
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute, "close")
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
self.assertEquals(51.0, volume_price)
def test_no_overwrite(self):
minute = self.market_opens[TEST_CALENDAR_START]
sid = 1
data = DataFrame(
data={
"open": [10.0],
"high": [20.0],
"low": [30.0],
"close": [40.0],
"volume": [50.0],
},
index=[minute],
)
self.writer.write_sid(sid, data)
with self.assertRaises(BcolzMinuteOverlappingData):
self.writer.write_sid(sid, data)
def test_append_to_same_day(self):
"""
Test writing data with the same date as existing data in our file.
"""
sid = 1
first_minute = self.market_opens[TEST_CALENDAR_START]
data = DataFrame(
data={
"open": [10.0],
"high": [20.0],
"low": [30.0],
"close": [40.0],
"volume": [50.0],
},
index=[first_minute],
)
self.writer.write_sid(sid, data)
# Write data in the same day as the previous minute
second_minute = first_minute + Timedelta(minutes=1)
new_data = DataFrame(
data={
"open": [5.0],
"high": [10.0],
"low": [3.0],
"close": [7.0],
"volume": [10.0],
},
index=[second_minute],
)
self.writer.write_sid(sid, new_data)
open_price = self.reader.get_value(sid, second_minute, "open")
self.assertEquals(5.0, open_price)
high_price = self.reader.get_value(sid, second_minute, "high")
self.assertEquals(10.0, high_price)
low_price = self.reader.get_value(sid, second_minute, "low")
self.assertEquals(3.0, low_price)
close_price = self.reader.get_value(sid, second_minute, "close")
self.assertEquals(7.0, close_price)
volume_price = self.reader.get_value(sid, second_minute, "volume")
self.assertEquals(10.0, volume_price)
def test_append_on_new_day(self):
sid = 1
ohlcv = {
"open": [2.0],
"high": [3.0],
"low": [1.0],
"close": [2.0],
"volume": [10.0],
}
dt = self.market_opens[TEST_CALENDAR_STOP]
data = DataFrame(data=ohlcv, index=[dt])
self.writer.write_sid(sid, data)
# Open a new writer to cover `open` method, also a common usage
# of appending new days will be writing to an existing directory.
cday = self.trading_calendar.schedule.index.freq
new_end_session = TEST_CALENDAR_STOP + cday
writer = BcolzMinuteBarWriter.open(self.dest, new_end_session)
next_day_minute = dt + cday
new_data = DataFrame(data=ohlcv, index=[next_day_minute])
writer.write_sid(sid, new_data)
# Get a new reader to test updated calendar.
reader = BcolzMinuteBarReader(self.dest)
second_minute = dt + Timedelta(minutes=1)
# The second minute should have been padded with zeros
for col in ("open", "high", "low", "close"):
assert_almost_equal(nan, reader.get_value(sid, second_minute, col))
self.assertEqual(0, reader.get_value(sid, second_minute, "volume"))
# The next day minute should have data.
for col in ("open", "high", "low", "close", "volume"):
assert_almost_equal(ohlcv[col],
reader.get_value(sid, next_day_minute, col))
def test_write_multiple_sids(self):
"""
Test writing multiple sids.
Tests both that the data is written to the correct sid, as well as
ensuring that the logic for creating the subdirectory path to each sid
does not cause issues from attempts to recreate existing paths.
(Calling out this coverage, because an assertion of that logic does not
show up in the test itself, but is exercised by the act of attempting
to write two consecutive sids, which would be written to the same
containing directory, `00/00/000001.bcolz` and `00/00/000002.bcolz)
Before applying a check to make sure the path writing did not
re-attempt directory creation an OSError like the following would
occur:
```
OSError: [Errno 17] File exists: '/tmp/tmpR7yzzT/minute_bars/00/00'
```
"""
minute = self.market_opens[TEST_CALENDAR_START]
sids = [1, 2]
data = DataFrame(
data={
"open": [15.0],
"high": [17.0],
"low": [11.0],
"close": [15.0],
"volume": [100.0],
},
index=[minute],
)
self.writer.write_sid(sids[0], data)
data = DataFrame(
data={
"open": [25.0],
"high": [27.0],
"low": [21.0],
"close": [25.0],
"volume": [200.0],
},
index=[minute],
)
self.writer.write_sid(sids[1], data)
sid = sids[0]
open_price = self.reader.get_value(sid, minute, "open")
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, "high")
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, "low")
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, "close")
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
self.assertEquals(100.0, volume_price)
sid = sids[1]
open_price = self.reader.get_value(sid, minute, "open")
self.assertEquals(25.0, open_price)
high_price = self.reader.get_value(sid, minute, "high")
self.assertEquals(27.0, high_price)
low_price = self.reader.get_value(sid, minute, "low")
self.assertEquals(21.0, low_price)
close_price = self.reader.get_value(sid, minute, "close")
self.assertEquals(25.0, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
self.assertEquals(200.0, volume_price)
def test_pad_data(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
day = TEST_CALENDAR_START + freq
minute = self.market_opens[day]
data = DataFrame(
data={
"open": [15.0],
"high": [17.0],
"low": [11.0],
"close": [15.0],
"volume": [100.0],
},
index=[minute],
)
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, "open")
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, "high")
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, "low")
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, "close")
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
self.assertEquals(100.0, volume_price)
# Check that if we then pad the rest of this day, we end up with
# 2 days worth of minutes.
self.writer.pad(sid, day)
self.assertEqual(
len(self.writer._ensure_ctable(sid)),
self.writer._minutes_per_day * 2,
)
def test_nans(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
minutes = date_range(minute, periods=9, freq="min")
data = DataFrame(
data={
"open": full(9, nan),
"high": full(9, nan),
"low": full(9, nan),
"close": full(9, nan),
"volume": full(9, 0.0),
},
index=minutes,
)
self.writer.write_sid(sid, data)
fields = ["open", "high", "low", "close", "volume"]
ohlcv_window = list(
map(
transpose,
self.reader.load_raw_arrays(
fields,
minutes[0],
minutes[-1],
[sid],
),
))
for i, field in enumerate(fields):
if field != "volume":
assert_array_equal(full(9, nan), ohlcv_window[i][0])
else:
assert_array_equal(zeros(9), ohlcv_window[i][0])
def test_differing_nans(self):
"""
Also test nans of differing values/construction.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
minutes = date_range(minute, periods=9, freq="min")
data = DataFrame(
data={
"open": ((0b11111111111 << 52) +
arange(1, 10, dtype=int64)).view(float64),
"high": ((0b11111111111 << 52) +
arange(11, 20, dtype=int64)).view(float64),
"low": ((0b11111111111 << 52) +
arange(21, 30, dtype=int64)).view(float64),
"close": ((0b11111111111 << 52) +
arange(31, 40, dtype=int64)).view(float64),
"volume":
full(9, 0.0),
},
index=minutes,
)
self.writer.write_sid(sid, data)
fields = ["open", "high", "low", "close", "volume"]
ohlcv_window = list(
map(
transpose,
self.reader.load_raw_arrays(
fields,
minutes[0],
minutes[-1],
[sid],
),
))
for i, field in enumerate(fields):
if field != "volume":
assert_array_equal(full(9, nan), ohlcv_window[i][0])
else:
assert_array_equal(zeros(9), ohlcv_window[i][0])
def test_write_cols(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
cols = {
"open": array([10.0, 11.0]),
"high": array([20.0, 21.0]),
"low": array([30.0, 31.0]),
"close": array([40.0, 41.0]),
"volume": array([50.0, 51.0]),
}
dts = array([minute_0, minute_1], dtype="datetime64[s]")
self.writer.write_cols(sid, dts, cols)
open_price = self.reader.get_value(sid, minute_0, "open")
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, "high")
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, "low")
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, "close")
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, "volume")
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, "open")
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, "high")
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, "low")
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, "close")
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, "volume")
self.assertEquals(51.0, volume_price)
def test_write_cols_mismatch_length(self):
dts = date_range(self.market_opens[self.test_calendar_start],
periods=2,
freq="min").asi8.astype("datetime64[s]")
sid = 1
cols = {
"open": array([10.0, 11.0, 12.0]),
"high": array([20.0, 21.0]),
"low": array([30.0, 31.0, 33.0, 34.0]),
"close": array([40.0, 41.0]),
"volume": array([50.0, 51.0, 52.0]),
}
with self.assertRaises(BcolzMinuteWriterColumnMismatch):
self.writer.write_cols(sid, dts, cols)
def test_unadjusted_minutes(self):
"""
Test unadjusted minutes.
"""
start_minute = self.market_opens[TEST_CALENDAR_START]
minutes = [
start_minute,
start_minute + Timedelta("1 min"),
start_minute + Timedelta("2 min"),
]
sids = [1, 2]
data_1 = DataFrame(
data={
"open": [15.0, nan, 15.1],
"high": [17.0, nan, 17.1],
"low": [11.0, nan, 11.1],
"close": [14.0, nan, 14.1],
"volume": [1000, 0, 1001],
},
index=minutes,
)
self.writer.write_sid(sids[0], data_1)
data_2 = DataFrame(
data={
"open": [25.0, nan, 25.1],
"high": [27.0, nan, 27.1],
"low": [21.0, nan, 21.1],
"close": [24.0, nan, 24.1],
"volume": [2000, 0, 2001],
},
index=minutes,
)
self.writer.write_sid(sids[1], data_2)
reader = BcolzMinuteBarReader(self.dest)
columns = ["open", "high", "low", "close", "volume"]
sids = [sids[0], sids[1]]
arrays = list(
map(
transpose,
reader.load_raw_arrays(
columns,
minutes[0],
minutes[-1],
sids,
),
))
data = {sids[0]: data_1, sids[1]: data_2}
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
assert_almost_equal(data[sid][col], arrays[i][j])
def test_unadjusted_minutes_early_close(self):
"""
Test unadjusted minute window, ensuring that early closes are filtered
out.
"""
day_before_thanksgiving = Timestamp("2015-11-25", tz="UTC")
xmas_eve = Timestamp("2015-12-24", tz="UTC")
market_day_after_xmas = Timestamp("2015-12-28", tz="UTC")
minutes = [
self.market_closes[day_before_thanksgiving] - Timedelta("2 min"),
self.market_closes[xmas_eve] - Timedelta("1 min"),
self.market_opens[market_day_after_xmas] + Timedelta("1 min"),
]
sids = [1, 2]
data_1 = DataFrame(
data={
"open": [15.0, 15.1, 15.2],
"high": [17.0, 17.1, 17.2],
"low": [11.0, 11.1, 11.3],
"close": [14.0, 14.1, 14.2],
"volume": [1000, 1001, 1002],
},
index=minutes,
)
self.writer.write_sid(sids[0], data_1)
data_2 = DataFrame(
data={
"open": [25.0, 25.1, 25.2],
"high": [27.0, 27.1, 27.2],
"low": [21.0, 21.1, 21.2],
"close": [24.0, 24.1, 24.2],
"volume": [2000, 2001, 2002],
},
index=minutes,
)
self.writer.write_sid(sids[1], data_2)
reader = BcolzMinuteBarReader(self.dest)
columns = ["open", "high", "low", "close", "volume"]
sids = [sids[0], sids[1]]
arrays = list(
map(
transpose,
reader.load_raw_arrays(
columns,
minutes[0],
minutes[-1],
sids,
),
))
data = {sids[0]: data_1, sids[1]: data_2}
start_minute_loc = self.trading_calendar.all_minutes.get_loc(
minutes[0])
minute_locs = [
self.trading_calendar.all_minutes.get_loc(minute) -
start_minute_loc for minute in minutes
]
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
assert_almost_equal(data[sid].loc[minutes, col],
arrays[i][j][minute_locs])
def test_adjust_non_trading_minutes(self):
start_day = Timestamp("2015-06-01", tz="UTC")
end_day = Timestamp("2015-06-02", tz="UTC")
sid = 1
cols = {
"open": arange(1, 781),
"high": arange(1, 781),
"low": arange(1, 781),
"close": arange(1, 781),
"volume": arange(1, 781),
}
dts = array(
self.trading_calendar.minutes_for_sessions_in_range(
self.trading_calendar.minute_to_session_label(start_day),
self.trading_calendar.minute_to_session_label(end_day),
))
self.writer.write_cols(sid, dts, cols)
self.assertEqual(
self.reader.get_value(sid,
Timestamp("2015-06-01 20:00:00", tz="UTC"),
"open"),
390,
)
self.assertEqual(
self.reader.get_value(sid,
Timestamp("2015-06-02 20:00:00", tz="UTC"),
"open"),
780,
)
with self.assertRaises(NoDataOnDate):
self.reader.get_value(sid, Timestamp("2015-06-02", tz="UTC"),
"open")
with self.assertRaises(NoDataOnDate):
self.reader.get_value(sid,
Timestamp("2015-06-02 20:01:00", tz="UTC"),
"open")
def test_adjust_non_trading_minutes_half_days(self):
# half day
start_day = Timestamp("2015-11-27", tz="UTC")
end_day = Timestamp("2015-11-30", tz="UTC")
sid = 1
cols = {
"open": arange(1, 601),
"high": arange(1, 601),
"low": arange(1, 601),
"close": arange(1, 601),
"volume": arange(1, 601),
}
dts = array(
self.trading_calendar.minutes_for_sessions_in_range(
self.trading_calendar.minute_to_session_label(start_day),
self.trading_calendar.minute_to_session_label(end_day),
))
self.writer.write_cols(sid, dts, cols)
self.assertEqual(
self.reader.get_value(sid,
Timestamp("2015-11-27 18:00:00", tz="UTC"),
"open"),
210,
)
self.assertEqual(
self.reader.get_value(sid,
Timestamp("2015-11-30 21:00:00", tz="UTC"),
"open"),
600,
)
self.assertEqual(
self.reader.get_value(sid,
Timestamp("2015-11-27 18:01:00", tz="UTC"),
"open"),
210,
)
with self.assertRaises(NoDataOnDate):
self.reader.get_value(sid, Timestamp("2015-11-30", tz="UTC"),
"open")
with self.assertRaises(NoDataOnDate):
self.reader.get_value(sid,
Timestamp("2015-11-30 21:01:00", tz="UTC"),
"open")
def test_set_sid_attrs(self):
"""Confirm that we can set the attributes of a sid's file correctly."""
sid = 1
start_day = Timestamp("2015-11-27", tz="UTC")
end_day = Timestamp("2015-06-02", tz="UTC")
attrs = {
"start_day": start_day.value / int(1e9),
"end_day": end_day.value / int(1e9),
"factor": 100,
}
# Write the attributes
self.writer.set_sid_attrs(sid, **attrs)
# Read the attributes
for k, v in attrs.items():
self.assertEqual(self.reader.get_sid_attr(sid, k), v)
def test_truncate_between_data_points(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + timedelta(days=1),
end=self.test_calendar_start + timedelta(days=3),
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
"open": [10.0, 11.0],
"high": [20.0, 21.0],
"low": [30.0, 31.0],
"close": [40.0, 41.0],
"volume": [50.0, 51.0],
},
index=minutes,
)
self.writer.write_sid(sid, data)
# Open a new writer to cover `open` method, also truncating only
# applies to an existing directory.
writer = BcolzMinuteBarWriter.open(self.dest)
# Truncate to first day with data.
writer.truncate(days[0])
# Refresh the reader since truncate update the metadata.
self.reader = BcolzMinuteBarReader(self.dest)
self.assertEqual(self.writer.last_date_in_output_for_sid(sid), days[0])
cal = self.trading_calendar
_, last_close = cal.open_and_close_for_session(days[0])
self.assertEqual(self.reader.last_available_dt, last_close)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, "open")
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, "high")
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, "low")
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, "close")
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, "volume")
self.assertEquals(50.0, volume_price)
def test_truncate_all_data_points(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(
start=self.test_calendar_start + timedelta(days=1),
end=self.test_calendar_start + timedelta(days=3),
)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(
data={
"open": [10.0, 11.0],
"high": [20.0, 21.0],
"low": [30.0, 31.0],
"close": [40.0, 41.0],
"volume": [50.0, 51.0],
},
index=minutes,
)
self.writer.write_sid(sid, data)
# Truncate to first day in the calendar, a day before the first
# day with minute data.
self.writer.truncate(self.test_calendar_start)
# Refresh the reader since truncate update the metadata.
self.reader = BcolzMinuteBarReader(self.dest)
self.assertEqual(
self.writer.last_date_in_output_for_sid(sid),
self.test_calendar_start,
)
cal = self.trading_calendar
_, last_close = cal.open_and_close_for_session(
self.test_calendar_start)
self.assertEqual(self.reader.last_available_dt, last_close)
def test_early_market_close(self):
# Date to test is 2015-11-30 9:31
# Early close is 2015-11-27 18:00
friday_after_tday = Timestamp("2015-11-27", tz="UTC")
friday_after_tday_close = self.market_closes[friday_after_tday]
before_early_close = friday_after_tday_close - timedelta(minutes=8)
after_early_close = friday_after_tday_close + timedelta(minutes=8)
monday_after_tday = Timestamp("2015-11-30", tz="UTC")
minute = self.market_opens[monday_after_tday]
# Test condition where there is data written after the market
# close (ideally, this should not occur in datasets, but guards
# against consumers of the minute bar writer, which do not filter
# out after close minutes.
minutes = [before_early_close, after_early_close, minute]
sid = 1
data = DataFrame(
data={
"open": [10.0, 11.0, nan],
"high": [20.0, 21.0, nan],
"low": [30.0, 31.0, nan],
"close": [40.0, 41.0, nan],
"volume": [50, 51, 0],
},
index=minutes,
)
self.writer.write_sid(sid, data)
open_price = self.reader.get_value(sid, minute, "open")
assert_almost_equal(nan, open_price)
high_price = self.reader.get_value(sid, minute, "high")
assert_almost_equal(nan, high_price)
low_price = self.reader.get_value(sid, minute, "low")
assert_almost_equal(nan, low_price)
close_price = self.reader.get_value(sid, minute, "close")
assert_almost_equal(nan, close_price)
volume = self.reader.get_value(sid, minute, "volume")
self.assertEquals(0, volume)
asset = self.asset_finder.retrieve_asset(sid)
last_traded_dt = self.reader.get_last_traded_dt(asset, minute)
self.assertEquals(
last_traded_dt,
before_early_close,
"The last traded dt should be before the early "
"close, even when data is written between the early "
"close and the next open.",
)
def test_minute_updates(self):
"""
Test minute updates.
"""
start_minute = self.market_opens[TEST_CALENDAR_START]
minutes = [
start_minute,
start_minute + Timedelta("1 min"),
start_minute + Timedelta("2 min"),
]
sids = [1, 2]
data_1 = DataFrame(
data={
"open": [15.0, nan, 15.1],
"high": [17.0, nan, 17.1],
"low": [11.0, nan, 11.1],
"close": [14.0, nan, 14.1],
"volume": [1000, 0, 1001],
},
index=minutes,
)
data_2 = DataFrame(
data={
"open": [25.0, nan, 25.1],
"high": [27.0, nan, 27.1],
"low": [21.0, nan, 21.1],
"close": [24.0, nan, 24.1],
"volume": [2000, 0, 2001],
},
index=minutes,
)
frames = {1: data_1, 2: data_2}
update_path = self.instance_tmpdir.getpath("updates.h5")
update_writer = H5MinuteBarUpdateWriter(update_path)
update_writer.write(frames)
update_reader = H5MinuteBarUpdateReader(update_path)
self.writer.write(update_reader.read(minutes, sids))
# Refresh the reader since truncate update the metadata.
reader = BcolzMinuteBarReader(self.dest)
columns = ["open", "high", "low", "close", "volume"]
sids = [sids[0], sids[1]]
arrays = list(
map(
transpose,
reader.load_raw_arrays(
columns,
minutes[0],
minutes[-1],
sids,
),
))
data = {sids[0]: data_1, sids[1]: data_2}
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
assert_almost_equal(data[sid][col], arrays[i][j])
class BcolzMinuteBarTestCase(TestCase):
@classmethod
def setUpClass(cls):
cls.env = TradingEnvironment()
all_market_opens = cls.env.open_and_closes.market_open
all_market_closes = cls.env.open_and_closes.market_close
indexer = all_market_opens.index.slice_indexer(
start=TEST_CALENDAR_START, end=TEST_CALENDAR_STOP)
cls.market_opens = all_market_opens[indexer]
cls.market_closes = all_market_closes[indexer]
cls.test_calendar_start = cls.market_opens.index[0]
cls.test_calendar_stop = cls.market_opens.index[-1]
def setUp(self):
self.dir_ = TempDirectory()
self.dir_.create()
self.dest = self.dir_.getpath('minute_bars')
os.makedirs(self.dest)
self.writer = BcolzMinuteBarWriter(
TEST_CALENDAR_START,
self.dest,
self.market_opens,
self.market_closes,
US_EQUITIES_MINUTES_PER_DAY,
)
self.reader = BcolzMinuteBarReader(self.dest)
def tearDown(self):
self.dir_.cleanup()
def test_write_one_ohlcv(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_two_bars(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
data = DataFrame(data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=[minute_0, minute_1])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_write_on_second_day(self):
second_day = self.test_calendar_start + 1
minute = self.market_opens[second_day]
sid = 1
data = DataFrame(data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
def test_write_empty(self):
minute = self.market_opens[self.test_calendar_start]
sid = 1
data = DataFrame(data={
'open': [0],
'high': [0],
'low': [0],
'close': [0],
'volume': [0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
assert_almost_equal(nan, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
assert_almost_equal(nan, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
assert_almost_equal(nan, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
assert_almost_equal(nan, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
assert_almost_equal(0, volume_price)
def test_write_on_multiple_days(self):
tds = self.market_opens.index
days = tds[tds.slice_indexer(start=self.test_calendar_start + 1,
end=self.test_calendar_start + 3)]
minutes = DatetimeIndex([
self.market_opens[days[0]] + timedelta(minutes=60),
self.market_opens[days[1]] + timedelta(minutes=120),
])
sid = 1
data = DataFrame(data={
'open': [10.0, 11.0],
'high': [20.0, 21.0],
'low': [30.0, 31.0],
'close': [40.0, 41.0],
'volume': [50.0, 51.0]
},
index=minutes)
self.writer.write(sid, data)
minute = minutes[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(50.0, volume_price)
minute = minutes[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(51.0, volume_price)
def test_no_overwrite(self):
minute = self.market_opens[TEST_CALENDAR_START]
sid = 1
data = DataFrame(data={
'open': [10.0],
'high': [20.0],
'low': [30.0],
'close': [40.0],
'volume': [50.0]
},
index=[minute])
self.writer.write(sid, data)
with self.assertRaises(BcolzMinuteOverlappingData):
self.writer.write(sid, data)
def test_write_multiple_sids(self):
"""
Test writing multiple sids.
Tests both that the data is written to the correct sid, as well as
ensuring that the logic for creating the subdirectory path to each sid
does not cause issues from attempts to recreate existing paths.
(Calling out this coverage, because an assertion of that logic does not
show up in the test itself, but is exercised by the act of attempting
to write two consecutive sids, which would be written to the same
containing directory, `00/00/000001.bcolz` and `00/00/000002.bcolz)
Before applying a check to make sure the path writing did not
re-attempt directory creation an OSError like the following would
occur:
```
OSError: [Errno 17] File exists: '/tmp/tmpR7yzzT/minute_bars/00/00'
```
"""
minute = self.market_opens[TEST_CALENDAR_START]
sids = [1, 2]
data = DataFrame(data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write(sids[0], data)
data = DataFrame(data={
'open': [25.0],
'high': [27.0],
'low': [21.0],
'close': [25.0],
'volume': [200.0]
},
index=[minute])
self.writer.write(sids[1], data)
sid = sids[0]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
sid = sids[1]
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(25.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(27.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(21.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(25.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(200.0, volume_price)
def test_pad_data(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
data = DataFrame(data={
'open': [15.0],
'high': [17.0],
'low': [11.0],
'close': [15.0],
'volume': [100.0]
},
index=[minute])
self.writer.write(sid, data)
open_price = self.reader.get_value(sid, minute, 'open')
self.assertEquals(15.0, open_price)
high_price = self.reader.get_value(sid, minute, 'high')
self.assertEquals(17.0, high_price)
low_price = self.reader.get_value(sid, minute, 'low')
self.assertEquals(11.0, low_price)
close_price = self.reader.get_value(sid, minute, 'close')
self.assertEquals(15.0, close_price)
volume_price = self.reader.get_value(sid, minute, 'volume')
self.assertEquals(100.0, volume_price)
def test_nans(self):
"""
Test writing empty data.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
minutes = date_range(minute, periods=9, freq='min')
data = DataFrame(data={
'open': full(9, nan),
'high': full(9, nan),
'low': full(9, nan),
'close': full(9, nan),
'volume': full(9, 0),
},
index=[minutes])
self.writer.write(sid, data)
fields = ['open', 'high', 'low', 'close', 'volume']
ohlcv_window = self.reader.unadjusted_window(fields, minutes[0],
minutes[-1], [sid])
for i, field in enumerate(fields):
if field != 'volume':
assert_array_equal(full(9, nan), ohlcv_window[i][0])
else:
assert_array_equal(zeros(9), ohlcv_window[i][0])
def test_differing_nans(self):
"""
Also test nans of differing values/construction.
"""
sid = 1
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertIs(last_date, NaT)
self.writer.pad(sid, TEST_CALENDAR_START)
last_date = self.writer.last_date_in_output_for_sid(sid)
self.assertEqual(last_date, TEST_CALENDAR_START)
freq = self.market_opens.index.freq
minute = self.market_opens[TEST_CALENDAR_START + freq]
minutes = date_range(minute, periods=9, freq='min')
data = DataFrame(data={
'open':
((0b11111111111 << 52) + arange(1, 10, dtype=int64)).view(float64),
'high': ((0b11111111111 << 52) +
arange(11, 20, dtype=int64)).view(float64),
'low': ((0b11111111111 << 52) +
arange(21, 30, dtype=int64)).view(float64),
'close': ((0b11111111111 << 52) +
arange(31, 40, dtype=int64)).view(float64),
'volume':
full(9, 0),
},
index=[minutes])
self.writer.write(sid, data)
fields = ['open', 'high', 'low', 'close', 'volume']
ohlcv_window = self.reader.unadjusted_window(fields, minutes[0],
minutes[-1], [sid])
for i, field in enumerate(fields):
if field != 'volume':
assert_array_equal(full(9, nan), ohlcv_window[i][0])
else:
assert_array_equal(zeros(9), ohlcv_window[i][0])
def test_write_cols(self):
minute_0 = self.market_opens[self.test_calendar_start]
minute_1 = minute_0 + timedelta(minutes=1)
sid = 1
cols = {
'open': array([10.0, 11.0]),
'high': array([20.0, 21.0]),
'low': array([30.0, 31.0]),
'close': array([40.0, 41.0]),
'volume': array([50.0, 51.0])
}
dts = array([minute_0, minute_1], dtype='datetime64[s]')
self.writer.write_cols(sid, dts, cols)
open_price = self.reader.get_value(sid, minute_0, 'open')
self.assertEquals(10.0, open_price)
high_price = self.reader.get_value(sid, minute_0, 'high')
self.assertEquals(20.0, high_price)
low_price = self.reader.get_value(sid, minute_0, 'low')
self.assertEquals(30.0, low_price)
close_price = self.reader.get_value(sid, minute_0, 'close')
self.assertEquals(40.0, close_price)
volume_price = self.reader.get_value(sid, minute_0, 'volume')
self.assertEquals(50.0, volume_price)
open_price = self.reader.get_value(sid, minute_1, 'open')
self.assertEquals(11.0, open_price)
high_price = self.reader.get_value(sid, minute_1, 'high')
self.assertEquals(21.0, high_price)
low_price = self.reader.get_value(sid, minute_1, 'low')
self.assertEquals(31.0, low_price)
close_price = self.reader.get_value(sid, minute_1, 'close')
self.assertEquals(41.0, close_price)
volume_price = self.reader.get_value(sid, minute_1, 'volume')
self.assertEquals(51.0, volume_price)
def test_write_cols_mismatch_length(self):
dts = date_range(self.market_opens[self.test_calendar_start],
periods=2,
freq='min').asi8.astype('datetime64[s]')
sid = 1
cols = {
'open': array([10.0, 11.0, 12.0]),
'high': array([20.0, 21.0]),
'low': array([30.0, 31.0, 33.0, 34.0]),
'close': array([40.0, 41.0]),
'volume': array([50.0, 51.0, 52.0])
}
with self.assertRaises(BcolzMinuteWriterColumnMismatch):
self.writer.write_cols(sid, dts, cols)
def test_unadjusted_minutes(self):
"""
Test unadjusted minutes.
"""
start_minute = self.market_opens[TEST_CALENDAR_START]
minutes = [
start_minute, start_minute + Timedelta('1 min'),
start_minute + Timedelta('2 min')
]
sids = [1, 2]
data_1 = DataFrame(data={
'open': [15.0, nan, 15.1],
'high': [17.0, nan, 17.1],
'low': [11.0, nan, 11.1],
'close': [14.0, nan, 14.1],
'volume': [1000, 0, 1001]
},
index=minutes)
self.writer.write(sids[0], data_1)
data_2 = DataFrame(data={
'open': [25.0, nan, 25.1],
'high': [27.0, nan, 27.1],
'low': [21.0, nan, 21.1],
'close': [24.0, nan, 24.1],
'volume': [2000, 0, 2001]
},
index=minutes)
self.writer.write(sids[1], data_2)
reader = BcolzMinuteBarReader(self.dest)
columns = ['open', 'high', 'low', 'close', 'volume']
sids = [sids[0], sids[1]]
arrays = reader.unadjusted_window(columns, minutes[0], minutes[-1],
sids)
data = {sids[0]: data_1, sids[1]: data_2}
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
assert_almost_equal(data[sid][col], arrays[i][j])
def test_unadjusted_minutes_early_close(self):
"""
Test unadjusted minute window, ensuring that early closes are filtered
out.
"""
day_before_thanksgiving = Timestamp('2015-11-25', tz='UTC')
xmas_eve = Timestamp('2015-12-24', tz='UTC')
market_day_after_xmas = Timestamp('2015-12-28', tz='UTC')
minutes = [
self.market_closes[day_before_thanksgiving] - Timedelta('2 min'),
self.market_closes[xmas_eve] - Timedelta('1 min'),
self.market_opens[market_day_after_xmas] + Timedelta('1 min')
]
sids = [1, 2]
data_1 = DataFrame(data={
'open': [15.0, 15.1, 15.2],
'high': [17.0, 17.1, 17.2],
'low': [11.0, 11.1, 11.3],
'close': [14.0, 14.1, 14.2],
'volume': [1000, 1001, 1002],
},
index=minutes)
self.writer.write(sids[0], data_1)
data_2 = DataFrame(data={
'open': [25.0, 25.1, 25.2],
'high': [27.0, 27.1, 27.2],
'low': [21.0, 21.1, 21.2],
'close': [24.0, 24.1, 24.2],
'volume': [2000, 2001, 2002],
},
index=minutes)
self.writer.write(sids[1], data_2)
reader = BcolzMinuteBarReader(self.dest)
columns = ['open', 'high', 'low', 'close', 'volume']
sids = [sids[0], sids[1]]
arrays = reader.unadjusted_window(columns, minutes[0], minutes[-1],
sids)
data = {sids[0]: data_1, sids[1]: data_2}
start_minute_loc = self.env.market_minutes.get_loc(minutes[0])
minute_locs = [
self.env.market_minutes.get_loc(minute) - start_minute_loc
for minute in minutes
]
for i, col in enumerate(columns):
for j, sid in enumerate(sids):
assert_almost_equal(data[sid].loc[minutes, col],
arrays[i][j][minute_locs])
def test_adjust_non_trading_minutes(self):
start_day = Timestamp('2015-06-01', tz='UTC')
end_day = Timestamp('2015-06-02', tz='UTC')
sid = 1
cols = {
'open': arange(1, 781),
'high': arange(1, 781),
'low': arange(1, 781),
'close': arange(1, 781),
'volume': arange(1, 781)
}
dts = array(self.env.minutes_for_days_in_range(start_day, end_day))
self.writer.write_cols(sid, dts, cols)
self.assertEqual(
self.reader.get_value(sid,
Timestamp('2015-06-01 20:00:00', tz='UTC'),
'open'), 390)
self.assertEqual(
self.reader.get_value(sid,
Timestamp('2015-06-02 20:00:00', tz='UTC'),
'open'), 780)
self.assertEqual(
self.reader.get_value(sid, Timestamp('2015-06-02', tz='UTC'),
'open'), 390)
self.assertEqual(
self.reader.get_value(sid,
Timestamp('2015-06-02 20:01:00', tz='UTC'),
'open'), 780)
def test_adjust_non_trading_minutes_half_days(self):
# half day
start_day = Timestamp('2015-11-27', tz='UTC')
end_day = Timestamp('2015-11-30', tz='UTC')
sid = 1
cols = {
'open': arange(1, 601),
'high': arange(1, 601),
'low': arange(1, 601),
'close': arange(1, 601),
'volume': arange(1, 601)
}
dts = array(self.env.minutes_for_days_in_range(start_day, end_day))
self.writer.write_cols(sid, dts, cols)
self.assertEqual(
self.reader.get_value(sid,
Timestamp('2015-11-27 18:00:00', tz='UTC'),
'open'), 210)
self.assertEqual(
self.reader.get_value(sid,
Timestamp('2015-11-30 21:00:00', tz='UTC'),
'open'), 600)
self.assertEqual(
self.reader.get_value(sid,
Timestamp('2015-11-27 18:01:00', tz='UTC'),
'open'), 210)
self.assertEqual(
self.reader.get_value(sid, Timestamp('2015-11-30', tz='UTC'),
'open'), 210)
self.assertEqual(
self.reader.get_value(sid,
Timestamp('2015-11-30 21:01:00', tz='UTC'),
'open'), 600)
