Analyze windows of time in a time series. You can preform the following operations:
- Slide a window of given length across your series
- Partion your series into a given number of windows
gem 'time_window', :git => 'https://github.com/olojac/time_window.git'You need a collection of objects with that respond to object.time. The collection should also be sorted from oldest to newest.
# your sorted collection of object with the attribute *#time*
list = [
OpenStruct.new(time: 3.hours.ago, value: 1),
OpenStruct.new(time: 2.hours.ago, value: 2),
OpenStruct.new(time: 1.hour.ago, value: 3),
]
# create a time series and slide a time window over it
TimeWindow.new(list, window_size: 1.hour).each do |items|
items # your sliding time window
endYou can also access the times and indexes for the window
TimeWindow.new(list, window_size: 1.hour).each do |items, times, indexes|
items # array of items in the window
times # array with start and end time of window
indexes # range of the indexes for the items in the window
endTimeWindow implements Enumerable so #map, #to_a and so on is also availible.
TimeWindow.slide is the default operation and the same as TimeWindow.new.
You can also set the step_size (default is 1.second).
TimeWindow.slide(list, window_size: 1.hour, step_size: 10.seconds).each do |items|
items # your sliding time window
endTimeWindow.partition divieds your list to a given number for equally large windows.
TimeWindow.partition(list, window_count: 10).each do |items|
items # your sliding time window
endThe gem is available as open source under the terms of the MIT License.