ThinkScript转Pinescript：消除复杂递归的实现方案问询

原ThinkScript代码 #

def state = {default init, long, short};
def trail;
def mclose;
switch (state[1]) {
case init:
   if (!IsNaN(loss)) {
      switch (firstTrade) { 
      case long:
         state = state.long; 
         mclose = close; 
         trail = close - loss;
      case short:
         state = state.short; 
         mclose = close; 
         trail = close + loss;
   }
} else {
   state = state.init;
   trail = Double.NaN; 
   mclose = Double.NaN;
}
case long:
   if (close > trail[1]) {
      state = state.long;
      mclose = Max(mclose[1], close); 
      trail = Max (trail[1], mclose - loss);
   } else {
      state = state.short; 
      mclose = close;
      trail = mclose + loss; 
   }
case short:
   if (close < trail[1]) {
      state = state.short;
      mclose = Min(mclose[1], close); 
      trail = Min(trail[1], mclose + loss);
   } else {
      state = state.long; 
      mclose = close; 
      trail = mclose - loss;
   } 
}

转换思路与PineScript实现 #

PineScript不支持递归，但可以通过var关键字声明持久化变量，模拟状态流转和变量的逐bar更新。核心是用var存储前一根bar的状态、trail和mclose值，然后在每根bar中根据前值计算当前值：

//@version=5
indicator("ThinkScript Port", overlay=true)

// 定义输入参数（需自行匹配原脚本的loss和firstTrade逻辑）
loss = input.float(1.0, "Loss Amount")
firstTrade = input.string("long", "First Trade Direction", options=["long", "short"])

// 用整数模拟状态枚举：0=init, 1=long, 2=short
var int state = 0
var float trail = na
var float mclose = na

// 逐bar计算当前状态、trail、mclose
if barstate.isfirst
    // 处理第一根bar的初始化
    if not na(loss)
        if firstTrade == "long"
            state := 1
            mclose := close
            trail := close - loss
        else
            state := 2
            mclose := close
            trail := close + loss
    else
        state := 0
        trail := na
        mclose := na
else
    switch state[1]
        case 0 // init状态逻辑
            if not na(loss)
                if firstTrade == "long"
                    state := 1
                    mclose := close
                    trail := close - loss
                else
                    state := 2
                    mclose := close
                    trail := close + loss
            else
                state := 0
                trail := na
                mclose := na
        case 1 // long状态逻辑
            if close > trail[1]
                state := 1
                mclose := math.max(mclose[1], close)
                trail := math.max(trail[1], mclose - loss)
            else
                state := 2
                mclose := close
                trail := mclose + loss
        case 2 // short状态逻辑
            if close < trail[1]
                state := 2
                mclose := math.min(mclose[1], close)
                trail := math.min(trail[1], mclose + loss)
            else
                state := 1
                mclose := close
                trail := mclose - loss

// 可选：可视化trail和当前状态
plot(trail, color=state == 1 ? color.green : color.red, title="Trail")
bgcolor(state == 1 ? color.new(color.green, 90) : state == 2 ? color.new(color.red, 90) : na)

关键说明 #

• 状态模拟：用整数替代ThinkScript的枚举类型，通过var声明确保状态在bar间持久化
• 变量更新：每根bar基于前一根的状态、trail、mclose值计算当前值，完全规避递归逻辑
• 边界处理：加入barstate.isfirst处理第一根bar的初始化，和原脚本逻辑完全对齐
• 输入适配：原脚本中的loss和firstTrade需根据实际场景替换为对应的PineScript输入或计算逻辑

内容的提问来源于stack exchange，提问作者Chris