Skip to content

Anterior 101

This guide will introduce you to the core concepts of Anterior and show you how to use them in your projects.

Backtesting

The BackTester is the main anterior class, serving as the base class to schedule functions to run live or in backtest simulations.

Schedule Functions

Write or attach simple deterministic schedules to your functions using the with any of five scheduling primitives: after, between, cron, every, and on.

Simple schedules 
from anterior.warp import BackTester

bt = BackTester()

@bt.after(seconds=5)
def decorated_func():
    print("I run after 5 seconds from the start of the backtest")

def declared_func():
    print("I run after 10 days from the start of the backtest")

bt.after(days=10).do(process.declared_func)

from anterior.warp import BackTester

bt = BackTester()

@bt.between(days_of_week="mon-fri")
def decorated_func():
    print("I run between Monday and Friday")

def declared_func():
    print("I run between January and June")

bt.between(months=(1, 6)).do(process.argument_func)

from anterior.warp import BackTester

bt = BackTester()

@schedules.cron("0 0 1 1 *")
def decorated_func():
    print("I run every year on January 1st")

def declared_func():
    print("I run every 5 minutes")

bt.cron("*/5 * * * *").do(process.declared_func)

from anterior.warp import BackTester

bt = BackTester()

@bt.cron("0 0 1 1 *")
def decorated_func():
    print("I run every 15 minutes")

def declared_func():
    print("I run every hour")

bt.cron("*/5 * * * *").do(process.argument_func)

from anterior.warp import BackTester

bt = BackTester()

@bt.on("2023-05-20 10:00")
def decorated_func():
    print("I run on May 20th, 2023 at 10 AM")

def declared_func():
    print("I run on January 1st, 2024 at midnight")

bt.on(year=2024, month=1, day=1).do(process.declared_func)

You can also or attach conditional schedules to your functions using the when and once methods. These methods allow you to run a function when or once a non-deterministic condition function is met.

Conditional schedules 
import random
from anterior.warp import BackTester

def conditional_func():
    return random.randint(0, 1)

def executable_func():
    print("I run when conditional_func returns 1")

bt = BackTester()
bt.when(conditional_func).do(executable_func)

import random
from anterior.warp import BackTester

def conditional_func():
    return random.randint(0, 1)

def executable_func():
    print("I run once when the conditional function returns 1, but only once")

bt = BackTester()
bt.once(conditional_func).do(executable_func)

Simple and conditional schedules can be combined using the & and | operators, representing logical AND and OR respectively.

Combined schedules 
from anterior.warp import BackTester

def conditional_func():
    return random.randint(0, 1)

def func1():
    print("I run between Monday and Friday when the conditional function returns 1")

def func2():
    print("I run every hour on January 1st, 2024")

bt = BackTester()
(bt.between(days_of_week="mon-fri") | bt.when(conditional_func)).do(func1)
(bt.on(year=2024, month=1, day=1) & bt.every(hours=1)).do(func2)

from anterior.warp import BackTester

def conditional_func():
    return random.randint(0, 1)

def func1():
    print("I run between 9:00 AM and 12:00 PM or when the conditional function returns 1")

def func2():
    print("I run on January 1st, 2024 or January 5th, 2024")

bt = BackTester()
(bt.between(hours=(9, 12)) | bt.when(conditional_func)).do(func1)
(bt.on(year=2024, month=1, day=1) | bt.on(year=2024, month=1, day=5)).do(func2)

Run live or backtest simulations

Once you have scheduled your functions, you can run them live and stop them at any time, or backtest and backfill them over a specified time period.

Run, Stop, Backtest & Backfill 
from anterior import BackTester

bt = MyBackTester()

counter = 0

@bt.every(hours=1)
def update():
    global counter
    counter += 1

bt.run()                                            # run a live process 
...
process.stop()                                      # stop the live process
process.run(start="2020-01-01", end="2024-01-01")   # specify start and end dates to backtest
process.run(start="2020-01-01")                     # specify only the start date to backfill

from anterior import BackTester

bt = MyBackTester()

counter = 0

@bt.every(hours=1)
def update():
    global counter
    counter += 1

bt.run()                                            # run a live process 
...
process.stop()                                      # stop the live process
process.run(start="2020-01-01", end="2024-01-01")   # specify start and end dates to backtest
process.run(start="2020-01-01")                     # specify only the start date to backfill

from anterior import BackTester

bt = MyBackTester()

counter = 0

@bt.every(hours=1)
def update():
    global counter
    counter += 1

bt.run()                                            # run a live process 
...
process.stop()                                      # stop the live process
process.run(start="2020-01-01", end="2024-01-01")   # specify start and end dates to backtest
process.run(start="2020-01-01")                     # specify only the start date to backfill

from anterior import BackTester

bt = MyBackTester()

counter = 0

@bt.every(hours=1)
def update():
    global counter
    counter += 1

bt.run()                                            # run a live process 
...
process.stop()                                      # stop the live process
process.run(start="2020-01-01", end="2024-01-01")   # specify start and end dates to backtest
process.run(start="2020-01-01")                     # specify only the start date to backfill

Oracle data

The OracleDataFrame and OracleSeries classes inherit all the methods and attributes of a pandas or polars DataFrame and Series, depending on the specification. However, they simplify historical data access by only returning data before the simulated time.

Oracle data usage 
import pandas as pd
from anterior import BackTester, OracleDataFrame

df = OracleDataFrame(pd.read_csv("data.csv"))   # data between 2020-01-01 and 2024-01-01

bt = BackTester()

@bt.do_every(hours=1)
def hour_update():
    df.iloc[-1]     # last value as of every simulated hour

@bt.do_every(days=1)
def day_update():
    df.iloc[-2:]    # last 2 value2 as of every simulated day

@bt.do_on("2023-01-01")
def day_update():
    df.iloc[-5:]    # last 5 values as of simulated January 1st, 2023

bt.run(start="2023-01-01", end="2024-01-01")

import pandas as pd
from anterior import BackTester, OracleSeries

df = OracleSeries(pd.read_csv("data.csv"))  # data between 2020-01-01 and 2024-01-01

bt = BackTester()

@bt.do_every(hours=1)
def hour_update():
    df.iloc[-1]     # last value as of every simulated hour

@bt.do_every(days=1)
def day_update():
    df.iloc[-1:2]   # last 2 value as of every simulated day

@bt.do_on("2023-01-01")
def day_update():
    df.iloc[-5:]    # last 5 values as of simulated January 1st, 2023

bt.run(start="2020-01-01", end="2024-01-01")