sortefi-ta

Technical indicators, candlestick patterns, chart patterns and price maths for pandas. Pure numpy + pandas: pip install sortefi-ta and go, no C library to compile.

Built and used in production by Sortefi, where you can describe a stock, crypto or forex screen in plain English and get the matching symbols.

Install

pip install sortefi-ta

Python 3.10+, numpy and pandas only.

Quick start

import pandas as pd
import sortefi_ta as ta

df = pd.read_csv("candles.csv")          # columns: open, high, low, close, volume (oldest first)

rsi = ta.calc_rsi(df, period=14)["rsi"]          # numpy array, one value per row
macd = ta.calc_macd(df)                           # {"macd", "signal", "histogram"}
pats = ta.calc_candlestick_patterns(df, ["hammer", "bullish_engulfing"])

print(rsi[-1])                    # latest RSI
print(pats["hammer"][-1] == 1.0)  # did the latest candle form a hammer?

Every function takes a DataFrame and returns a dict of numpy arrays aligned with the rows, where [-1] is the latest candle. Warm-up rows are NaN. ta.INDICATOR_FUNCTIONS lists them all.

What's inside

Notes

• Candlestick patterns are shape-only. A hammer is a hammer whatever came before it; combine with a trend tool (moving-average slope, ADX, calc_stage) when you need "hammer after a downtrend".
• Every threshold is in the source. Each pattern lives in its own short file with its rule written out, e.g. candlestick/hammer.py.
• Numbers follow common charting conventions (EMA seeded with an SMA, Wilder smoothing for RSI/ATR/ADX, population standard deviation for Bollinger Bands).

Tests

pip install -e ".[test]" && pytest

License

MIT — see LICENSE and NOTICE.