ClickHouse/tests/queries/0_stateless/01558_ttest_scipy.python

#!/usr/bin/env python3
import os
import sys
from scipy import stats
import pandas as pd
import numpy as np

CURDIR = os.path.dirname(os.path.realpath(__file__))
sys.path.insert(0, os.path.join(CURDIR, "helpers"))

from pure_http_client import ClickHouseClient


def test_and_check(name, a, b, t_stat, p_value, precision=1e-2):
    client = ClickHouseClient()
    client.query("DROP TABLE IF EXISTS ttest;")
    client.query("CREATE TABLE ttest (left Float64, right UInt8) ENGINE = Memory;")
    client.query(
        "INSERT INTO ttest VALUES {};".format(
            ", ".join(["({},{})".format(i, 0) for i in a])
        )
    )
    client.query(
        "INSERT INTO ttest VALUES {};".format(
            ", ".join(["({},{})".format(j, 1) for j in b])
        )
    )
    real = client.query_return_df(
        "SELECT roundBankers({}(left, right).1, 16) as t_stat, ".format(name)
        + "roundBankers({}(left, right).2, 16) as p_value ".format(name)
        + "FROM ttest FORMAT TabSeparatedWithNames;"
    )
    real_t_stat = real["t_stat"][0]
    real_p_value = real["p_value"][0]
    assert (
        abs(real_t_stat - np.float64(t_stat)) < precision
    ), "clickhouse_t_stat {}, scipy_t_stat {}".format(real_t_stat, t_stat)
    assert (
        abs(real_p_value - np.float64(p_value)) < precision
    ), "clickhouse_p_value {}, scipy_p_value {}".format(real_p_value, p_value)
    client.query("DROP TABLE IF EXISTS ttest;")


def test_student():
    rvs1 = np.round(stats.norm.rvs(loc=1, scale=5, size=500), 2)
    rvs2 = np.round(stats.norm.rvs(loc=10, scale=5, size=500), 2)
    s, p = stats.ttest_ind(rvs1, rvs2, equal_var=True)
    test_and_check("studentTTest", rvs1, rvs2, s, p)

    rvs1 = np.round(stats.norm.rvs(loc=0, scale=5, size=500), 2)
    rvs2 = np.round(stats.norm.rvs(loc=0, scale=5, size=500), 2)
    s, p = stats.ttest_ind(rvs1, rvs2, equal_var=True)
    test_and_check("studentTTest", rvs1, rvs2, s, p)

    rvs1 = np.round(stats.norm.rvs(loc=2, scale=10, size=512), 2)
    rvs2 = np.round(stats.norm.rvs(loc=5, scale=20, size=1024), 2)
    s, p = stats.ttest_ind(rvs1, rvs2, equal_var=True)
    test_and_check("studentTTest", rvs1, rvs2, s, p)

    rvs1 = np.round(stats.norm.rvs(loc=0, scale=10, size=1024), 2)
    rvs2 = np.round(stats.norm.rvs(loc=0, scale=10, size=512), 2)
    s, p = stats.ttest_ind(rvs1, rvs2, equal_var=True)
    test_and_check("studentTTest", rvs1, rvs2, s, p)


def test_welch():
    rvs1 = np.round(stats.norm.rvs(loc=1, scale=15, size=500), 2)
    rvs2 = np.round(stats.norm.rvs(loc=10, scale=5, size=500), 2)
    s, p = stats.ttest_ind(rvs1, rvs2, equal_var=False)
    test_and_check("welchTTest", rvs1, rvs2, s, p)

    rvs1 = np.round(stats.norm.rvs(loc=0, scale=7, size=500), 2)
    rvs2 = np.round(stats.norm.rvs(loc=0, scale=3, size=500), 2)
    s, p = stats.ttest_ind(rvs1, rvs2, equal_var=False)
    test_and_check("welchTTest", rvs1, rvs2, s, p)

    rvs1 = np.round(stats.norm.rvs(loc=0, scale=10, size=1024), 2)
    rvs2 = np.round(stats.norm.rvs(loc=5, scale=1, size=512), 2)
    s, p = stats.ttest_ind(rvs1, rvs2, equal_var=False)
    test_and_check("welchTTest", rvs1, rvs2, s, p)

    rvs1 = np.round(stats.norm.rvs(loc=5, scale=10, size=512), 2)
    rvs2 = np.round(stats.norm.rvs(loc=5, scale=10, size=1024), 2)
    s, p = stats.ttest_ind(rvs1, rvs2, equal_var=False)
    test_and_check("welchTTest", rvs1, rvs2, s, p)


if __name__ == "__main__":
    test_student()
    test_welch()
    print("Ok.")