pylan-lib 0.2.0

Python library that simulates the combined impact of recurring events.

Pylan is a Python library that simulates the impact of scheduled events. You can install the Python library using PyPi with the following command:

pip install pylan-lib

This code snippet shows some basic functionality when doing simulations.

from pylan import Item, Subtract, Add, Multiply

savings = Item(start_value=100)
salary_payments = Add("1m", 2500, offset="24d") # Salary paid every month at the 24th
salary_increase = Multiply("1y", 1.2) # Salary grows each year 20%
mortgage = Subtract("0 0 2 * *", 1500)  # cron support

salary_payments.add_pattern(salary_increase) # Add increase to salary pattern
savings.add_patterns([salary_payments, mortgage])
result = savings.run("2024-1-1", "2028-1-1")

x, y = result.plot_axes()

plt.plot(x, y)
plt.show()

There are 2 important classes in this library: Item and Pattern. A pattern is an abstract base class, with multiple implementations. These implementations resemble a time based pattern (e.g. add 10 every month, yearly inflation, etc). The Item is something that patterns can be added to, like a savings account.

---

Class: Granularity

Refers to the minimum step size needed for iterations given a set of patterns. Can be tweaked for Item.run(). Note that the default value here is the minimum granularity of the added patterns. Supports: hour, day, week, month, year

>>> from pylan import Granularity
>>> savings.run("2024-1-1", "2028-1-1", Granularity.day)
>>> savings.run("2024-1-1", "2028-1-1", Granularity.month)

---

Class: Item

An item that you can apply patterns to and simulate over time. Optionally, you can set a start value.

>>> savings = Item(start_value=100)

Item.add_pattern(self, pattern: Pattern) -> None:

Add a pattern object to this item.

>>> test = Add(["2024-1-4", "2024-2-1"], 1)
>>> savings = Item(start_value=100)
>>> savings.add_pattern(test)

Item.add_patterns(self, patterns: list[Pattern]) -> None:

Adds a list of patterns object to this item.

>>> gains = Multiply("4m", 1)
>>> adds = Multiply("2d", 1)
>>> savings = Item(start_value=100)
>>> savings.add_patterns([gains, adds])

Item.run(

Runs the provided patterns between the start and end date. Creates a result object with all the iterations per day/month/etc.

>>> savings = Item(start_value=100)
>>> savings.add_patterns([gains, adds])
>>> savings.run("2024-1-1", "2025-1-1")

Item.until(

Runs the provided patterns until a stop value is reached. Returns the timedelta needed to reach the stop value. NOTE: Don't use offset with a start date here.

>>> savings = Item(start_value=100)
>>> savings.add_patterns([gains, adds])
>>> savings.until(200)  # returns timedelta

Item.iterate(

Creates Iterator object for the item. Can be used in a for loop. Returns a tuple of datetime and item object.

>>> for date, saved in savings.iterate("2024-1-1", "2025-2-2", Granularity.day):
>>>     print(date, saved.value)

---

Class: Result

Outputted by an item run. Result of a simulation between start and end date. Has the schedule and values as attributes (which are both lists).

>>> result = savings.run("2024-1-1", "2024-3-1")
>>> x, y = result.plot_axes() # can be used for matplotlib
>>> result.final # last value
>>> result.to_csv("test.csv")

Result.str(self) -> str:

String format of result is a column oriented table with dates and values.

Result.repr(self) -> str:

String format of result is a column oriented table with dates and values.

Result.getitem(self, key: str | datetime) -> float | int:

Get a result by the date using a dict key.

>>> print(result["2024-5-5"])

Result.final(self):

Returns the result on the last day of the simulation.

>>> result = savings.run("2024-1-1", "2024-3-1")
>>> result.final

Result.valid(self):

Returns true if the result has a valid format

Result.plot_axes(self, categorical_x_axis: bool = False) -> tuple[list, list]:

Returns x, y axes of the simulated run. X axis are dates and Y axis are values.

>>> result = savings.run("2024-1-1", "2024-3-1")
>>> x, y = result.plot_axes() # can be used for matplotlib

Result.to_csv(self, filename: str, sep: str = ";") -> None:

Exports the result to a csv file. Row oriented.

>>> result = savings.run("2024-1-1", "2024-3-1")
>>> result.to_csv("test.csv")

---

Class: Pattern

Pattern is an abstract base class with the following implementations:

• Add(schedule, value)
• Subtract(schedule, value)
• Multiply(schedule, value)
• Divide(schedule, value)

Note, all implementations have the following optional parameters:

• start_date: str or datetime with the minimum date for the pattern to start
• end_date: str or datetime, max date for the pattern
• offset: str, offsets each occurence of the pattern based on the start date

>>> mortgage = Subtract("0 0 2 * *", 1500)  # cron support
>>> inflation = Divide(["2025-1-1", "2026-1-1", "2027-1-1"], 1.08)

Pattern.apply(self) -> None:

Applies the pattern to the item provided as a parameter. Implemented in the specific classes.

Pattern.add_pattern(self, pattern: Any) -> None:

Applies the pattern to the value of this pattern. E.g. You add a salary each month, over time this salary can grow using another pattern.

Pattern.scheduled(self, current: datetime) -> bool:

Returns true if pattern is scheduled on the provided date.

---

Schedule

Passed to patterns as a parameter. Is converted to a list of datetime objects. Accepts multiple formats.

Cron schedules

For example, "0 0 2 * *" runs on the second day of each month.

Timedelta strings

Combination of a count and timedelta. For example, 2d (every 2 days) 3m (every 3 months). Currently supports: years (y), months (m), weeks (w), days (d).

Timedelta lists

Same as timedelta, but then alternates between the schedules. For example, ["2d", "5d"] will be triggered after 2 days, then after 5 days, then after 2 days, etc...

Datetime lists

A list of datetime objects or str that resemble datetime objects. For example, ["2024-1-1", "2025-1-1"].

NOTE: The date format in pylan is yyyy-mm-dd. Currently this is not configurable.