def test_resample():
df = to_dataframe(data)
assert isinstance(resample(df, "2d"), DataFrame)
assert list(resample(df, "2d").index.values[-2:]) == [
numpy.datetime64("2019-05-05T00:00:00.000000000"),
numpy.datetime64("2019-05-07T00:00:00.000000000"),
]
def test_resample_calendar():
df = to_dataframe(data)
assert isinstance(resample(df, "W-Mon"), DataFrame)
assert list(resample(df, "W-Mon").index.values[-2:]) == [
numpy.datetime64("2019-05-06T00:00:00.000000000"),
numpy.datetime64("2019-05-13T00:00:00.000000000"),
]
def resample_ohlc(price, vol, scale, mode='ohlcv'):
timescales = {
'5min': 5,
'15min': 15,
'30min': 30,
'1h': 60,
'4h': 240,
'12h': 720,
'1d': 1440,
'3d': 4320,
'1w': 10080
}
t = timescales.get(scale)
if mode == 'ohlcv':
new_price = resample(price, scale)
new_vol = list(new_price.volume)
if mode == 'close':
new_price = []
new_vol = []
for i in range(int(len(price) / t)):
pos = i * t
p = price[pos + t - 1]
new_price.append(p)
v = sum(vol[pos:pos + t])
new_vol.append(v)
if mode == 'mean':
new_price = []
new_vol = []
for i in range(int(len(price) / t)):
pos = i * t
p = statistics.mean(price[pos:pos + t])
new_price.append(p)
v = sum(vol[pos:pos + t])
new_vol.append(v)
if mode == 'median':
new_price = []
new_vol = []
for i in range(int(len(price) / t)):
pos = i * t
p = statistics.median(price[pos:pos + t])
new_price.append(p)
v = sum(vol[pos:pos + t])
new_vol.append(v)
return new_price, new_vol
def resample(ohlc: DataFrame, interval: str) -> DataFrame:
"""resample by <interval>"""
return resample(ohlc, interval)