AZPytest

A small Python backtesting toolkit: replay tick data from a CSV one row at a time, place simulated long/short trades, and run price analytics.

Setup

Requires Python 3.10+.

Install from PyPI:

pip install azpytest

Or straight from GitHub:

pip install git+https://github.com/vxvira/AZPytest.git

To work on it from a clone, install it editable along with pytest:

pip install -e . pytest

Usage

import AZPytest.handling as dh

cfg = dh.Configuration()
cfg.path         = "data.csv"
cfg.tickSz       = 0.25          # price increment of one tick
cfg.tickVal      = 0.25          # money value of one tick, per contract
cfg.priceAlias   = "price"       # CSV column names
cfg.tsEventAlias = "ts_event"
cfg.tsRecvAlias  = "ts_recv"
cfg.actionAlias  = "action"
cfg.buildConfig()

handler = dh.Handling(cfg)

handler.tick()                   # read the next row
handler.addLong(1)               # or addShort(n); adds change the net position
handler.tick()
handler.unrealizedPnl()          # pnl of the open trade
handler.closeLong()              # or closeShort()
handler.realizedPnl()            # pnl of the most recently closed trade

addLong/addShort adjust the net position: the same side grows it, the opposite side shrinks it, reaching zero ends the trade, and crossing zero flips it into a new trade. The first opposite-side add logs a warning.

Every tick is recorded in handler.data (price, tsEvent, tsRecv, action lists) and handler.positionsData ([inTrade, unrealizedPnl, entryIdx] per tick).

Analytics

AZPytest/analytics has two classes that work on pandas Series:

PriceAnalytics(PriceSeries, VolumeSeries=None)

Volume-dependent methods raise [fatal] if no VolumeSeries was given.

SetAnalytics(Series)

import pandas as pd
from AZPytest.analytics.PriceAnalytics import PriceAnalytics

pa = PriceAnalytics(pd.Series(handler.data.price))
pa.returnExponentialMovingAverage(20)
pa.returnRelativeStrengthIndex(14)

Tests

python -m pytest

License

MIT, see LICENSE.