This module contains some simple methods that are useful when analysing time
series data (as arrays time and values).
Building only on numpy, they reliably let you do things like:
- Find the indices strictly corresponding to a point in time, or the start and end times of a range (including end points, half-open, etc)
- Find the index closest to a point not appearing in the data, and interpolate the value at that point.
- Perform some action (max, min, mean, absolute max etc) on a range
- Check if a series is regularly spaced, or strictly non-decreasing, and obtain the sampling time
- Smooth using sliding windows, gaussian blur, and others
- Get an amplitude or power spectrum of periodic data
In short, things you can easily do with numpy, but without having to remember the syntax, and with unit tests for that particular operation. For a full overview, see the docs.
The code is tested on a recent version of Ubuntu & Python 3, but is so simple that it should work everywhere else too.
To install the latest release from PyPI, use
pip install datkit
To install from the repo, use e.g.
python setup.py install -e .
Tests can then be run with
python -m unittest
And docs can be built with
cd docs
make clean html