finkit

Open-source financial analysis library. Production-grade indicators, signals, risk metrics, screening, and cost analysis.

No VecTrade account required — this is a standalone library for the quant/fintech community.

Installation

pip install vectrade-finkit

Quick Start

import pandas as pd
import finkit

# ── Technical Indicators ──
df["sma_20"] = finkit.sma(df["close"], period=20)
df["ema_12"] = finkit.ema(df["close"], period=12)
df["rsi"] = finkit.rsi(df["close"], period=14)
macd_line, signal, histogram = finkit.macd(df["close"])
upper, middle, lower = finkit.bollinger_bands(df["close"])
df["atr"] = finkit.atr(df["high"], df["low"], df["close"], period=14)
df["vwap"] = finkit.vwap(df["high"], df["low"], df["close"], df["volume"])
df["obv"] = finkit.obv(df["close"], df["volume"])

# ── Signal Detection ──
df["buy_signal"] = finkit.crossover(df["sma_10"], df["sma_50"])
df["sell_signal"] = finkit.crossunder(df["sma_10"], df["sma_50"])
df["divergence"] = finkit.divergence(df["close"], df["rsi"])

# ── Signal Engine (composable rules) ──
engine = finkit.SignalEngine()
engine.add_rule("rsi_oversold", lambda df: df["rsi"] < 30, direction="long")
engine.add_rule("macd_cross", lambda df: finkit.crossover(df["macd"], df["signal"]), direction="long")
signals = engine.evaluate(df)

# ── Risk Metrics ──
sharpe = finkit.sharpe_ratio(returns, risk_free_rate=0.04)
sortino = finkit.sortino_ratio(returns)
mdd = finkit.max_drawdown(equity_curve)
value_at_risk = finkit.var(returns, confidence=0.95)

# ── Stock Screening ──
from finkit import Rule, screen

results = screen(universe_df, rules=[
    Rule("pe_ratio", "<", 25),
    Rule("market_cap", ">", 10_000_000_000),
    Rule("rsi_14", "between", (30, 70)),
    Rule("sector", "in", ["Technology", "Healthcare"]),
])

# ── Cost Analysis ──
cost = finkit.calculate_trade_cost(shares=100, price=150.0, commission_per_share=0.005)
drag = finkit.annual_cost_drag(trades_per_year=200, avg_trade_cost=cost.total, portfolio_value=100_000)

API Reference

finkit.indicators

finkit.signals

finkit.risk

finkit.screen

Supported operators: <, <=, >, >=, ==, !=, between, in, contains

finkit.costs

Design Principles

• Zero API dependency — works with any pandas DataFrame
• NumPy vectorized — fast computation on large datasets
• Minimal dependencies — only numpy and pandas
• Fully typed — py.typed marker, works with mypy/pyright
• Well-tested — 99% branch coverage

Part of the VecTrade Ecosystem

Documentation

Full documentation is available at docs.vectrade.io/sdks/finkit.

Contributing

See CONTRIBUTING.md for development setup and guidelines.

License

Apache-2.0 — see LICENSE.