使用 Numba JIT 编译创建高性能自定义技术指标 - 金融量化分析工具

🇺🇸English

Create a custom technical indicator with Numba JIT compilation for production-grade speed.

Arguments

• $0 = indicator name (e.g., zscore, squeeze, vwap-bands, custom-rsi, mean-reversion). Required.

If no arguments, ask the user what indicator they want to build.

Instructions

1. Read the indicator-expert rules, especially:
   • rules/custom-indicators.md — Numba patterns and templates
   • rules/numba-optimization.md — Performance best practices
   • rules/indicator-catalog.md — Check if indicator already exists in openalgo.ta
2. Check first : If the indicator already exists in openalgo.ta, tell the user and show the existing API
3. Create custom_indicators/{indicator_name}/ directory (on-demand)
4. Create {indicator_name}.py with:

File Structure

"""
{Indicator Name} — Custom Indicator
Description: {what it measures}
Category: {trend/momentum/volatility/volume/oscillator}
"""
import numpy as np
from numba import njit
import pandas as pd

# --- Core Computation (Numba JIT) ---
@njit(cache=True, nogil=True)
def _compute_{name}(data: np.ndarray, period: int) -> np.ndarray:
    """Numba-compiled core computation."""
    n = len(data)
    result = np.full(n, np.nan)
    # ... O(n) algorithm ...
    return result

# --- Public API ---
def {name}(data, period=20):
    """
    {Indicator Name}

    Args:
        data: Close prices (numpy array, pandas Series, or list)
        period: Lookback period (default: 20)

    Returns:
        Same type as input with indicator values
    """
    if isinstance(data, pd.Series):
        idx = data.index
        result = _compute_{name}(data.values.astype(np.float64), period)
        return pd.Series(result, index=idx, name="{Name}({period})")
    arr = np.asarray(data, dtype=np.float64)
    return _compute_{name}(arr, period)

5. Create chart.py for visualization:

"""Chart the custom indicator with Plotly."""
import os
from pathlib import Path
from datetime import datetime, timedelta
from dotenv import find_dotenv, load_dotenv
from openalgo import api, ta
import plotly.graph_objects as go
from plotly.subplots import make_subplots
from {indicator_name} import {name}

# ... fetch data, compute indicator, create chart ...

6. Create benchmark.py for performance testing:

"""Benchmark the custom indicator."""
import numpy as np
import time
from {indicator_name} import {name}

# Warmup
data = np.random.randn(1000)
_ = {name}(data, 20)

# Benchmark on different sizes
for size in [10_000, 100_000, 500_000]:
    data = np.random.randn(size)
    t0 = time.perf_counter()
    _ = {name}(data, 20)
    elapsed = (time.perf_counter() - t0) * 1000
    print(f"{size:>10,} bars: {elapsed:>8.2f}ms")

Numba Rules (CRITICAL)

MUST DO

• @njit(cache=True, nogil=True) on all compute functions
• np.full(n, np.nan) to initialize output arrays
• Use np.isnan() for NaN checks
• Explicit for loops (Numba compiles to machine code)
• O(n) algorithms: rolling sum, EMA recursion, deque-based extrema
• Float64 for all numeric arrays

MUST NOT

• Never fastmath=True (breaks np.isnan())
• Never use pandas inside @njit
• Never use try/except, dicts, sets, strings inside @njit
• Never call non-jitted functions from inside @njit

Available Building Blocks

These existing functions can be called inside @njit:

from openalgo.indicators.utils import (
    sma, ema, ema_wilder, stdev, true_range, atr_wilder,
    highest, lowest, rolling_sum, crossover, crossunder
)

Common Custom Indicator Patterns

Example Usage

/custom-indicator zscore /custom-indicator squeeze-momentum /custom-indicator vwap-bands /custom-indicator range-filter

Weekly Installs

91

Repository

marketcalls/ope…r-skills

GitHub Stars

5

First Seen

Feb 28, 2026

Security Audits

Gen Agent Trust HubWarnSocketPassSnykPass

Installed on

opencode91

codex91

gemini-cli90

github-copilot89

amp89

cline89