_{supertrend(high, low, close, length=10, mult=3.0) and rsi(close, length=14) on synthetic data · generated with matplotlib by docs/assets/make_example_chart.py}

The scope is deliberate: indicator math only. No charting, no broker integrations, no strategy execution framework.

Contents

• Installation
• Quick start
• Correctness guarantees
• What is not verified
• Conventions
• API layout
• Development

Installation

Install from PyPI:

pip install v1indicators

Or install from source for development:

git clone https://github.com/Vatthu/v1indicators.git
cd v1indicators
pip install -e ".[dev]"

Requires Python 3.10+, numpy, pandas, numba.

Quick start

import pandas as pd
from v1indicators import rsi, macd, supertrend

df = pd.read_csv("data.csv")

# single output -> named Series
df["RSI_14"] = rsi(df["close"], length=14)

# multiple outputs -> DataFrame with stable column names
macd_df = macd(df["close"], fast=12, slow=26, signal=9)
st_df = supertrend(df["high"], df["low"], df["close"], length=10, mult=3.0)

Every output preserves the input index and length.

Correctness guarantees

Enforced by tests and CI (Python 3.10 / 3.12 / 3.14). A regression in any of them fails the build.

What is not verified

⚠️ Warning: Six indicators compose the primitives above into trading signals. Their outputs are causal by construction — that is all this library claims about them. Nothing here measures whether their signals predict anything.

Validate on your own data before trading any of them.

Conventions

API layout

Import from the root:

from v1indicators import ema, sma, rsi, atr, obv

or from a family module:

from v1indicators.momentum import rsi, stoch

A second namespace mirrors how indicators are built:

from v1indicators.foundational.overlap import ema   # direct from price/volume
from v1indicators.derived.trend import supertrend   # built on other indicators

Family modules re-export both layers and remain the compatibility surface.

ℹ️ Note: New work incubates in experimental/ in the repository, which is not shipped in the binary PyPI package.

Most indicators take close only; range-based ones take high/low/close; volume indicators add volume. Series must share an index.

Development

pytest                                     # full suite (~7 s)
pytest -q tests/test_causality.py          # prefix invariance, all functions
pytest -q tests/test_causality_params.py   # same, non-default parameters
pytest -q tests/test_parity_core.py        # textbook reference parity
pytest -q tests/test_interoperability_matrix.py

🛡️ Guarantees: The causality gates discover public functions automatically: a new repainting indicator fails the suite without a test being written for it.

See CHANGELOG.md for full release history. License: MIT.