A python signal indicator code running on quantdinger platform

[brokermr810]

`my_indicator_name = "SuperTrend"
my_indicator_description = "SuperTrend (TradingView v4 logic). Chart outputs buy/sell only; backend normalizes execution."

df = df.copy()

Parameters (match Pine defaults)

periods = 10
multiplier = 3.0
change_atr = True # True: Wilder ATR (TradingView atr), False: SMA(TR)

Source = hl2

src = (df["high"] + df["low"]) / 2.0

True Range (TradingView tr)

prev_close = df["close"].shift(1)
tr0 = df["high"] - df["low"]
tr1 = (df["high"] - prev_close).abs()
tr2 = (df["low"] - prev_close).abs()
tr = pd.concat([tr0, tr1, tr2], axis=1).max(axis=1)

n = len(df)
close_arr = df["close"].values

--- ATR: TradingView-style Wilder ATR (seed with SMA, then recursive) ---

atr_sma = tr.rolling(periods).mean()

if change_atr:
atr = pd.Series([np.nan] * n, index=df.index, dtype="float64")
if n >= periods:
atr.iloc[periods - 1] = float(tr.iloc[:periods].mean())
for i in range(periods, n):
atr.iloc[i] = (atr.iloc[i - 1] * (periods - 1) + float(tr.iloc[i])) / periods
else:
atr = atr_sma

--- SuperTrend bands with Pine := recursion ---

up_final = [np.nan] * n
dn_final = [np.nan] * n

for i in range(n):
a = atr.iloc[i]
if pd.isna(a):
continue

up_raw = float(src.iloc[i] - multiplier * a)
dn_raw = float(src.iloc[i] + multiplier * a)

if i == 0 or pd.isna(up_final[i - 1]):
    up1 = up_raw
else:
    up1 = float(up_final[i - 1])

if i == 0 or pd.isna(dn_final[i - 1]):
    dn1 = dn_raw
else:
    dn1 = float(dn_final[i - 1])

prev_c = float(close_arr[i - 1]) if i > 0 else float(close_arr[i])

# up := close[1] > up1 ? max(up, up1) : up
up_final[i] = max(up_raw, up1) if prev_c > up1 else up_raw
# dn := close[1] < dn1 ? min(dn, dn1) : dn
dn_final[i] = min(dn_raw, dn1) if prev_c < dn1 else dn_raw

--- Trend recursion (Pine trend := ...) ---

trend = [1] * n
for i in range(1, n):
if pd.isna(up_final[i - 1]) or pd.isna(dn_final[i - 1]):
trend[i] = trend[i - 1]
continue

prev_t = trend[i - 1]
if prev_t == -1 and float(close_arr[i]) > float(dn_final[i - 1]):
    trend[i] = 1
elif prev_t == 1 and float(close_arr[i]) < float(up_final[i - 1]):
    trend[i] = -1
else:
    trend[i] = prev_t

trend_s = pd.Series(trend, index=df.index)

Plots (linebreak effect using None)

up_plot = [float(up_final[i]) if (trend[i] == 1 and not pd.isna(up_final[i])) else None for i in range(n)]
dn_plot = [float(dn_final[i]) if (trend[i] == -1 and not pd.isna(dn_final[i])) else None for i in range(n)]

Signals (exact Pine conditions)

buy_signal = (trend_s == 1) & (trend_s.shift(1) == -1)
sell_signal = (trend_s == -1) & (trend_s.shift(1) == 1)
buy_signal = buy_signal.fillna(False).astype(bool)
sell_signal = sell_signal.fillna(False).astype(bool)

Required for backend normalization (execution)

df["buy"] = buy_signal
df["sell"] = sell_signal

Chart marks: Pine uses plotshape(buySignal ? up : na) and plotshape(sellSignal ? dn : na)

buy_marks = [float(up_final[i]) if (bool(buy_signal.iloc[i]) and not pd.isna(up_final[i])) else None for i in range(n)]
sell_marks = [float(dn_final[i]) if (bool(sell_signal.iloc[i]) and not pd.isna(dn_final[i])) else None for i in range(n)]

output = {
"name": my_indicator_name,
"plots": [
{"name": "Up Trend", "type": "line", "data": up_plot, "color": "#00C853", "overlay": True},
{"name": "Down Trend", "type": "line", "data": dn_plot, "color": "#D50000", "overlay": True},
],
"signals": [
{"type": "buy", "text": "Buy", "data": buy_marks, "color": "#00E676"},
{"type": "sell", "text": "Sell", "data": sell_marks, "color": "#FF5252"},
],
}`