eventlogic 0.2.0

A lightweight library for performing logical operations on event-style or timestamp data.

Description

EventLogic is a lightweight library for performing logical operations on event-style or timestamp data. Logical operations on event-style data are commonplace in many fields. This library seeks to provide a generalized framework and methods for dealing with this data.

Installation

EventLogic can be installed with pip.

pip install --upgrade pip
pip install eventlogic

Common Usage

The most common way to use EventLogic is through the Events class, which handles multiple events simultaneously.

Creating Events

from eventlogic import Event, Events

#Create from list of events
events1 = Events([Event(1.0, 2.5), Event(2.9, 3.0), Event(5.0, 7.0)])

# Create directly from numpy arrays (recommended)
import numpy as np
ons = np.array([1, 3, 5])
offs = np.array([2, 4, 6])
events2 = Events.from_arrays(ons, offs)

# creating from datetime arrays
dates_on = np.array(['2023-01-01', '2023-01-02'], dtype='datetime64[D]')
dates_off = np.array(['2023-01-02', '2023-01-03'], dtype='datetime64[D]')
datetime_events = Events.from_arrays(dates_on, dates_off)

Logical Operations

#Intersection (find overlapping periods)
intersection = events1 & events2
print(intersection) # Shows all overlapping time periods

#Union (combine events, merging overlaps)
union = events1 | events2
print(union) # Shows combined events with overlaps merged

#Check if events are contained within other events
events1 = Events([Event(1, 5), Event(7, 10)])
events2 = Events([Event(2, 4), Event(8, 9)])
contained = events1.contains_events(events2) # Returns boolean array

Merging & Filtering

#Merge events that are close together
merged = events1.merge(threshold=5) # Merges events less than 5 units apart

#Filter events by duration
filtered = events1.duration_filter(lower_bound=0.5, upper_bound=2.0)

Examples

There are 9 flavors of event interactions:

Case	Events
1
2
3
4
5
6
7
8
9

Usage

The Event class represents an event with a start (on) and end (off) time. It supports various logical operations:

Operation Class	Symbols
Comparison Operators	>, <, >=, <=, ==, !=
Logical Operators	&, |, ^
Containment	in, not in

Creating an Event

from eventlogic import Event
event = Event(on=1, off=5)

Event Duration

duration = event.duration()  # Returns the duration of the event

Copying an Event

event_copy = event.copy()

Event Existence

exists = event.exists()  # Checks if the event is defined

Comparison Operators

If we look at case 1:

from eventlogic import Event
a = Event(3,4)
b = Event(1,2)
a > b # True
a < b # False

Logical Operations

event1 = Event(on=1, off=5)
event2 = Event(on=4, off=6)

Intersection

intersection = event1 & event2  # Returns (4,5)

Union

union = event1 | event2  # Returns [(1,6)]

xor

xor = event1 ^ event2  # Returns ((1,4), (5,6))

Containment Operators

If we look at case 9:

from eventlogic import Event
a = Event(3,4)
b = Event(2,5)
a in b # True
a not in b # False

Working with numpy.datetime64

import numpy as np
from eventlogic import Event

a = Event(np.datetime64('2023-01-01T12:00'), np.datetime64('2023-01-01T14:00'))
b = Event(np.datetime64('2023-01-01T13:00'), np.datetime64('2023-01-01T15:00'))

print(a & b)  # (2023-01-01T13:00,2023-01-01T14:00)
print(a | b)  # [(2023-01-01T12:00,2023-01-01T15:00)]
print(a > b)  # False
print(a < b)  # False

Event Container

The Events class allows for handling multiple events at once.

from eventlogic import Event, Events
events = Events([Event(1, 2), Event(3, 4), Event(5, 6)])
print(len(events))  # 3
print(events)  # [(1,2), (3,4), (5,6)]

And importantly, the creation of Events from numpy arrays (including datetime!).

import numpy as np
from eventlogic import Events

ons = np.array([1,3,5])
offs = np.array([2,4,6])
events = Events.from_arrays(ons,offs)
print(len(events))  # 3
print(events)  # [(1,2), (3,4), (5,6)]

Merging Events

merged_events = events.merge(threshold=0.5)
print(merged_events)

Filtering Events by Duration

filtered_events = events.duration_filter(lower_bound=0.5, upper_bound=2.0)
print(filtered_events)