Python
from rpy2.robjects import pandas2ri
import rpy2.robjects as ro
from rpy2.robjects.packages import importr

print('两独立样本率优效性检验的样本量计算 - Python程序 \nby Lokyshin\n')
# 请注意，该程序是python调用R的pwr.2p.test / pwr.p2n2.test与R内置的power.prop.test实现。
# reference: https://cran.r-project.org/web/packages/pwr

p1=0.5
p2=0.8
alpha=0.05
power=0.90
power_cal_method="binomial_enumeration"
R=2
alternative="two.sided"

# 打开pandas转r dataframe的功能
#pandas2ri.activate()
# 载入需要的R包
pwr = importr('pwr')

# 设定R代码
if R == 1 :
    if power_cal_method == "binomial_enumeration":
        #使用二项枚举方法计算
        r_codes =f'''

            # 计算效应量
            h <- ES.h({p2}, {p1})

            # 使用 if 语句根据 alternative 参数进行条件判断
            if ("{alternative}" == "two.sided") {{
                # H0: p2 = p1, Ha: p2 != p1
                sample_size <- pwr.2p.test(h = h, sig.level = {alpha}, power = {power}, alternative = "two.sided")
            }} else {{
                if ({p2} > {p1}) {{
                    #H0: p2 <= p1, Ha: p2 > p1
                    sample_size <- pwr.2p.test(h = h, sig.level = {alpha}, power ={power}, alternative = "greater")
                }} else {{
                    #H0: p2 >= p1, Ha: p2 < p1
                    sample_size <- pwr.2p.test(h = h, sig.level = {alpha}, power = {power}, alternative = "less")
                }}
            }}
            

            # 打印结果
            n <- ceiling(sample_size$n)
            total_sample_size <- n * 2
            R_result_str = paste0("使用二项枚举方法计算：",
                "试验组和对照组样本量均为: ", n, "；",
                "总样本量为: ", total_sample_size, "。")
            
            print(R_result_str)
            print(sample_size)

        '''
    else:
        #使用正态分布近似方法计算
        r_codes = f'''

            # 计算样本量
            result <- power.prop.test(p1 = {p1}, p2 = {p2}, sig.level = {alpha}, power = {power}, alternative = "{alternative}")

            # 调整结果
            n <- ceiling(result$n)
            total_sample_size <- n * 2

                # 打印结果
            R_result_str = paste0("使用二项枚举方法计算：",
                "试验组和对照组样本量均为: ", n, "；",
                "总样本量为: ", total_sample_size, "。")
            print(R_result_str)
            print(result)
                
        '''
else:
    ratio = R
    if power_cal_method == "binomial_enumeration":
        #使用二项枚举方法迭代计算
        r_codes =f'''

            # 计算效应量
            h <- ES.h({p2}, {p1})

            # 初始化迭代计数器
            iteration_counter <<- 0

            # 定义计算样本量的函数
            sample_size_function <- function(n1) {{
                iteration_counter <<- iteration_counter + 1  # 递增迭代计数器
                n2 <- {ratio} * n1
                
                # 使用 if 语句根据 alternative 参数进行条件判断
                if ("{alternative}" == "two.sided") {{
                    # H0: p2 = p1, Ha: p2 != p1
                    result <- pwr.2p2n.test(h = h, n1 = n1, n2 = n2, sig.level = {alpha}, power = NULL, alternative = "two.sided")
                }} else {{
                    if ({p2} > {p1}) {{
                        #H0: p2 <= p1, Ha: p2 > p1
                        result <- pwr.2p2n.test(h = h, n1 = n1, n2 = n2, sig.level = {alpha}, power = NULL, alternative = "greater")
                    }} else {{
                        #H0: p2 >= p1, Ha: p2 < p1
                        result <- pwr.2p2n.test(h = h, n1 = n1, n2 = n2, sig.level = {alpha}, power = NULL, alternative = "less")
                    }}
                }}

                cat("迭代次数:", iteration_counter, "； n1:", n1, "； n2:", n2, "\n")
                print(result)
                return(result$power - {power})
            }}

            # 使用 uniroot 函数进行迭代计算
            result <- uniroot(sample_size_function, interval = c(10, 10000))

            # 得到样本量
            n1 <- ceiling(result$root)
            n2 <- ceiling({ratio} * n1)

            R_result_str = paste0("使用二项枚举方法迭代计算：", 
                "对照组样本量为: ", n1, "；",
                "试验组样本量为: ", n2, "；", 
                "总样本量为: ", n1+n2, "。")
            
            print(R_result_str)

        '''
    else:
        #使用正态分布近似方法迭代计算
        power_target = power

        r_codes = f'''

            # 初始化迭代计数器
            iteration_counter <- 0

            if ("{alternative}" == "one.sided"){{
                cat('进行单侧检验:\n')
            }}else{{
                cat('进行双侧检验:\n')
            }}

            # 迭代计算样本量
            for (n1 in 1:10000) {{
                n2 <- {ratio} * n1
                total_n <- n1 + n2
                
                # 计算合并比例
                p_combined <- ({p1} * n1 +{p2} * n2) / (n1 + n2)
                
                # 计算标准误
                se <- sqrt(p_combined * (1 - p_combined) * (1 / n1 + 1 / n2))
                
                # 计算Z值
                if ("{alternative}" == "one.sided") {{
                    z_alpha <- qnorm(1 - {alpha}) # 单侧检验
                }} else {{
                    z_alpha <- qnorm(1 - {alpha} / 2)  # 双侧检验需要将alpha减半
                }}
                
                z_effect <- ({p2} - {p1}) / se
                
                # 计算功效power
                if ("{alternative}" == "one.sided") {{
                    power <- pnorm(z_effect - z_alpha)
                }} else {{
                    power <- pnorm(z_effect - z_alpha) + (1 - pnorm(z_effect + z_alpha))
                }}
                
                iteration_counter <- iteration_counter + 1
                cat("迭代次数:", iteration_counter, "； n1:", n1, "； n2:", n2, "； 实际Power:", power, "\n")
                
                if (power >={ power_target}) {{
                    break
                }}
            }}

            R_result_str = paste0("使用正态分布近似方法迭代计算：", 
                "对照组样本量为: ", n1, "；",
                "试验组样本量为: ", n2, "；", 
                "总样本量为: ", n1+n2, "。")
            print(R_result_str)

        '''
        
# 执行R代码
ro.r(r_codes)