crx-indicators 1.3.0

Standard indicators library for the Coinrule X execution platform

Coinrule X Indicators

Standard library of trading indicators for the Coinrule X execution platform.

Installation

pip install crx-indicators

Usage Patterns

Coinrule X Indicators supports two main calculation patterns depending on your environment:

1. Stateless (Batch) Calculation

Ideal for backtesting or one-off analysis where you have a full history of candles.

indicator = RSI(period=14)
result = indicator.calculate(candles) # Processes entire history

2. Stateful (Incremental) Calculation

Optimized for live trading and minimized latency. Instead of recalculating the entire history on every tick, the indicator maintains internal state and updates in $O(1)$ time.

# Initialization (typically on startup)
indicator = RSI(period=14)
indicator.seed(historical_candles)

# Live update (on every new closed candle)
latest_value = indicator.update(new_candle)

Structure

Indicators are located in the coinrule_x_indicators/indicators directory. Each indicator implements the standard interface CandleIndicator defined in coinrule_x_indicators.core.

Contributing

Git hooks (release version)

This repo ships a pre-commit hook that bumps the semver minor in pyproject.toml (e.g. 1.2.0 → 1.3.0) when you commit changes under coinrule_x_indicators/ or to pyproject.toml itself, but the version line was left unchanged from HEAD. If you already bumped the version in the same commit, the hook does nothing.

One-time setup (from the git repository root — monorepo root or this repo’s root):

bash crx-indicators/scripts/install-git-hooks.sh

For a standalone clone of crx-indicators only:

bash scripts/install-git-hooks.sh

This sets git config core.hooksPath to the git-hooks directory next to pyproject.toml. If your team uses another core.hooksPath, merge this pre-commit into that system instead of overwriting the config.

We welcome contributions of new indicators! Please follow these guidelines to ensure consistency and quality.

Adding a New Indicator

1. Create the Indicator File: Create a new file in coinrule_x_indicators/indicators/ (e.g., ema.py).

2. Implement the Class: Inherit from the CandleIndicator base class and implement both the calculate (stateless) and update (stateful) methods.

   from coinrule_x_indicators.core import CandleIndicator, Candle
   from typing import List, Union
   
   class EMA(CandleIndicator):
       def __init__(self, period: int = 8):
           self.period = period
           self.alpha = 2.0 / (period + 1.0)
           super().__init__(period=period)
   
       def calculate(self, candles: List[Candle]) -> float:
           # Batch implementation (e.g., using pandas)
           pass
   
       def update(self, candle: Candle) -> float:
           # Incremental implementation (O(1) update)
           # Update self._value and return it
           pass
   

3. Register the Indicator: Add your new class to coinrule_x_indicators/indicators/__init__.py to export it.

4. Update Registry: Add your indicator to coinrule_x_indicators/registry.yaml. This is used by the Coinrule X platform to discover available indicators.

   ema:
     latest: "1.0.0"
     versions:
       "1.0.0":
         class: "coinrule_x_indicators.indicators.ema.EMA"
         label: "EMA"
         description: "Exponential Moving Average"
         arguments:
           period:
             type: int
             default: 8
             label: "Period"
             description: "Number of periods for EMA calculation"
   

5. Validate Registry: Ensure your registry entry is valid before submitting:

   # Run validation script
   poetry run python coinrule_x_indicators/validation.py coinrule_x_indicators/registry.yaml
   

Versioning & Immutability Policy

To guarantee strategy reproducibility, Coinrule X Indicators follows a strict Add-Only policy for breaking changes. Strategies written today must produce the exact same signals 5 years from now.

1. Immutable Logic

Once an indicator version (e.g., ema v1.0.0) is published and used, its calculation logic MUST NOT change.

• Bug Fixes: Critical bugs in logic require a Patch version (e.g., v1.0.1).
• New Features: New arguments or logic require a New Version (e.g., v1.1.0 or v2.0.0).

2. Creating New Versions

If you need to change the behavior of an indicator:

1. Do NOT edit definitions of existing released classes.
2. Create a New Class: e.g., class EMAv2(Indicator).
3. Register New Version: Add the new version to registry.yaml alongside the old one.

Example registry.yaml structure for multi-version support:

ema:
  latest: "2.0.0"
  label: "EMA"
  versions:
    "1.0.0":
      class: "coinrule_x_indicators.indicators.ema.EMA"
      description: "Standard Exponential Moving Average"
    "2.0.0":
      class: "coinrule_x_indicators.indicators.ema_v2.EMAv2"
      description: "EMA with adjustable smoothing factor"
      arguments:
        period: { label: "Period", type: int, default: 14 }
        smoothing: { label: "Smoothing", type: float, default: 2.0 }

3. Deprecation

Old versions remain in the codebase indefinitely unless they pose a security risk. They can be marked as deprecated: true in the registry to warn developers against using them for new strategies, but existing strategies will continue to function.

Testing Guidelines

Every indicator must have a corresponding test file in the tests/ directory.

1. Create Test File: Create tests/test_<indicator_name>.py (e.g., tests/test_ema.py).

2. Use Standard Data: Use the provided load_candles utility to load the standardized test dataset (tests/data/candles.json). This dataset contains HYPE 1h candles, with the latest candle starting at 24 Dec 2025 15:00 UTC. This ensures all indicators are tested against the same market conditions.

   from tests.utils import load_candles
   from coinrule_x_indicators.indicators.ema import EMA
   
   def test_ema_calculation():
       candles = load_candles("candles.json")
       ema = EMA(period=8)
       result = ema.calculate(candles)
   
       # Verify result against known valid value (e.g. from TradingView or known lib)
       assert round(result, 2) == <EXPECTED_VALUE>
   

3. Required Test Cases:

   • Initialization: Verify arguments are stored correctly.
   • Batch Calculation: Verify that calculate() correctly processes a list of candles.
   • Precise Value: A test asserting the exact value (to 2 decimal places) against the provided dataset.
   • Incremental Consistency: Verify that update() produces the same result as calculate() when processing the same sequence of candles.

Running Tests

Run the test suite using pytest:

poetry run pytest

Submission Process

1. Fork the Repository: Create a fork of the repository to your own GitHub account.
2. Create a Feature Branch: Create a new branch for your changes (e.g., feat/add-rsi-indicator).
3. Implement & Test: Apply your changes and ensure all tests pass.
4. Create Pull Request: Submit a pull request to the main repository.

Note: Version numbers and the official CHANGELOG will be updated by maintainers upon release.

Development Guidelines

• Type Hinting: All methods must be fully type-hinted.
• Pandas/Numpy: Use vectorized operations where possible for performance.
• Zero Dependencies: Do not add new external dependencies without prior discussion.