timeseries

Overview

timeseries is a Python package that provides a one-dimensional time series data structure built entirely from the Python Standard Library. Time series objects can be instantiated from CSV files and provide basic computational capabilities, including sample statistics, overloaded arithmetic operations and filtering (rolling windows, exponential moving averages etc.).

Documentation

The latest build of the documentation is available on GitHub Pages. Documentation is generated with sphinx using the Read the Docs theme.

Installation

Python version 3.7 at minimum is required for the package, since the logic depends on ordering of dictionaries and datetime functionality introduced in version 3.7. There are no third-party module dependencies. The package can be installed directly with the command

$ pip3 install git+https://github.com/austerj/timeseries

Alternatively the package can be cloned and installed locally in editable mode with

$ git clone https://github.com/austerj/timeseries
$ pip3 install -e ./timeseries

or even imported directly from the REPL in the parent directory with

import timeseries as ts

Running unit tests

The package uses the doctest and unittest modules from the Python Standard Library for testing. All tests can be executed by running the command

$ python3 -m unittest -v

in the parent directory of the cloned repository.