Skip to content

CLN: test_moments_rolling.py for quantile/kurt/skew #36784

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 7 commits into from
Oct 2, 2020
Merged

CLN: test_moments_rolling.py for quantile/kurt/skew #36784

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
83dc354
Add new file for kurt and skew
Sep 30, 2020
b011fba
add tests
Sep 30, 2020
8aab8a2
Merge remote-tracking branch 'upstream/master' into clean/moments_rol…
Oct 1, 2020
8ed6a6a
Fix min_periods argument
Oct 1, 2020
3c85fa2
Port over quantile tests
Oct 1, 2020
c637db2
Remove unused import'
Oct 1, 2020
12abb02
Add missing no scipy decorator
Oct 1, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
226 changes: 1 addition & 225 deletions pandas/tests/window/moments/test_moments_rolling.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,187 +1,12 @@
import numpy as np
from numpy.random import randn
import pytest

import pandas.util._test_decorators as td

import pandas as pd
from pandas import DataFrame, Series, isna, notna
from pandas import DataFrame, Series
import pandas._testing as tm

import pandas.tseries.offsets as offsets


def _check_moment_func(
static_comp,
name,
raw,
has_min_periods=True,
has_center=True,
has_time_rule=True,
fill_value=None,
zero_min_periods_equal=True,
series=None,
frame=None,
**kwargs,
):
def get_result(obj, window, min_periods=None, center=False):
r = obj.rolling(window=window, min_periods=min_periods, center=center)
return getattr(r, name)(**kwargs)

series_result = get_result(series, window=50)
assert isinstance(series_result, Series)
tm.assert_almost_equal(series_result.iloc[-1], static_comp(series[-50:]))

frame_result = get_result(frame, window=50)
assert isinstance(frame_result, DataFrame)
tm.assert_series_equal(
frame_result.iloc[-1, :],
frame.iloc[-50:, :].apply(static_comp, axis=0, raw=raw),
check_names=False,
)

# check time_rule works
if has_time_rule:
win = 25
minp = 10
ser = series[::2].resample("B").mean()
frm = frame[::2].resample("B").mean()

if has_min_periods:
series_result = get_result(ser, window=win, min_periods=minp)
frame_result = get_result(frm, window=win, min_periods=minp)
else:
series_result = get_result(ser, window=win, min_periods=0)
frame_result = get_result(frm, window=win, min_periods=0)

last_date = series_result.index[-1]
prev_date = last_date - 24 * offsets.BDay()

trunc_series = series[::2].truncate(prev_date, last_date)
trunc_frame = frame[::2].truncate(prev_date, last_date)

tm.assert_almost_equal(series_result[-1], static_comp(trunc_series))

tm.assert_series_equal(
frame_result.xs(last_date),
trunc_frame.apply(static_comp, raw=raw),
check_names=False,
)

# excluding NaNs correctly
obj = Series(randn(50))
obj[:10] = np.NaN
obj[-10:] = np.NaN
if has_min_periods:
result = get_result(obj, 50, min_periods=30)
tm.assert_almost_equal(result.iloc[-1], static_comp(obj[10:-10]))

# min_periods is working correctly
result = get_result(obj, 20, min_periods=15)
assert isna(result.iloc[23])
assert not isna(result.iloc[24])

assert not isna(result.iloc[-6])
assert isna(result.iloc[-5])

obj2 = Series(randn(20))
result = get_result(obj2, 10, min_periods=5)
assert isna(result.iloc[3])
assert notna(result.iloc[4])

if zero_min_periods_equal:
# min_periods=0 may be equivalent to min_periods=1
result0 = get_result(obj, 20, min_periods=0)
result1 = get_result(obj, 20, min_periods=1)
tm.assert_almost_equal(result0, result1)
else:
result = get_result(obj, 50)
tm.assert_almost_equal(result.iloc[-1], static_comp(obj[10:-10]))

# window larger than series length (#7297)
if has_min_periods:
for minp in (0, len(series) - 1, len(series)):
result = get_result(series, len(series) + 1, min_periods=minp)
expected = get_result(series, len(series), min_periods=minp)
nan_mask = isna(result)
tm.assert_series_equal(nan_mask, isna(expected))

nan_mask = ~nan_mask
tm.assert_almost_equal(result[nan_mask], expected[nan_mask])
else:
result = get_result(series, len(series) + 1, min_periods=0)
expected = get_result(series, len(series), min_periods=0)
nan_mask = isna(result)
tm.assert_series_equal(nan_mask, isna(expected))

nan_mask = ~nan_mask
tm.assert_almost_equal(result[nan_mask], expected[nan_mask])

# check center=True
if has_center:
if has_min_periods:
result = get_result(obj, 20, min_periods=15, center=True)
expected = get_result(
pd.concat([obj, Series([np.NaN] * 9)]), 20, min_periods=15
)[9:].reset_index(drop=True)
else:
result = get_result(obj, 20, min_periods=0, center=True)
print(result)
expected = get_result(
pd.concat([obj, Series([np.NaN] * 9)]), 20, min_periods=0
)[9:].reset_index(drop=True)

tm.assert_series_equal(result, expected)

# shifter index
s = [f"x{x:d}" for x in range(12)]

if has_min_periods:
minp = 10

series_xp = (
get_result(
series.reindex(list(series.index) + s), window=25, min_periods=minp
)
.shift(-12)
.reindex(series.index)
)
frame_xp = (
get_result(
frame.reindex(list(frame.index) + s), window=25, min_periods=minp
)
.shift(-12)
.reindex(frame.index)
)

series_rs = get_result(series, window=25, min_periods=minp, center=True)
frame_rs = get_result(frame, window=25, min_periods=minp, center=True)

else:
series_xp = (
get_result(
series.reindex(list(series.index) + s), window=25, min_periods=0
)
.shift(-12)
.reindex(series.index)
)
frame_xp = (
get_result(
frame.reindex(list(frame.index) + s), window=25, min_periods=0
)
.shift(-12)
.reindex(frame.index)
)

series_rs = get_result(series, window=25, min_periods=0, center=True)
frame_rs = get_result(frame, window=25, min_periods=0, center=True)

if fill_value is not None:
series_xp = series_xp.fillna(fill_value)
frame_xp = frame_xp.fillna(fill_value)
tm.assert_series_equal(series_xp, series_rs)
tm.assert_frame_equal(frame_xp, frame_rs)


def test_centered_axis_validation():

Expand Down Expand Up @@ -716,33 +541,6 @@ def test_rolling_max_min_periods():
pd.Series([1, 2, 3]).rolling(window=3, min_periods=5).max()


@pytest.mark.parametrize("q", [0.0, 0.1, 0.5, 0.9, 1.0])
def test_rolling_quantile(q, raw, series, frame):
def scoreatpercentile(a, per):
values = np.sort(a, axis=0)

idx = int(per / 1.0 * (values.shape[0] - 1))

if idx == values.shape[0] - 1:
retval = values[-1]

else:
qlow = float(idx) / float(values.shape[0] - 1)
qhig = float(idx + 1) / float(values.shape[0] - 1)
vlow = values[idx]
vhig = values[idx + 1]
retval = vlow + (vhig - vlow) * (per - qlow) / (qhig - qlow)

return retval

def quantile_func(x):
return scoreatpercentile(x, q)

_check_moment_func(
quantile_func, name="quantile", quantile=q, raw=raw, series=series, frame=frame
)


def test_rolling_quantile_np_percentile():
# #9413: Tests that rolling window's quantile default behavior
# is analogous to Numpy's percentile
Expand Down Expand Up @@ -845,25 +643,3 @@ def test_rolling_std_neg_sqrt():

b = a.ewm(span=3).std()
assert np.isfinite(b[2:]).all()


@td.skip_if_no_scipy
def test_rolling_skew(raw, series, frame):
from scipy.stats import skew

_check_moment_func(
lambda x: skew(x, bias=False), name="skew", raw=raw, series=series, frame=frame
)


@td.skip_if_no_scipy
def test_rolling_kurt(raw, series, frame):
from scipy.stats import kurtosis

_check_moment_func(
lambda x: kurtosis(x, bias=False),
name="kurt",
raw=raw,
series=series,
frame=frame,
)
Loading