A python library for calculating Allan deviation and related time & frequency statistics. GPL v3+ license.

Developed at https://github.com/aewallin/allantools and also available on PyPi at https://pypi.python.org/pypi/AllanTools Discussion group at https://groups.google.com/d/forum/allantools

Input data should be evenly spaced observations of either fractional frequency, or phase in seconds. Deviations are calculated for given tau values in seconds.

These statistics are currently included:

• adev() Allan deviation
• oadev() overlapping Allan deviation,
• mdev() modified Allan deviation,
• tdev() Time deviation
• hdev() Hadamard deviation
• ohdev() overlapping Hadamard deviation
• totdev() total Allan deviation
• mtie() Maximum time interval error
• tierms() Time interval error RMS
• mtotdev() Modified total deviation
• ttotdev() Time total deviation
• htotdev() Hadamard total deviation
• theo1() Thêo1 deviation

Noise generators for creating synthetic datasets are also included:

• violet noise with f^2 PSD
• white noise with f^0 PSD
• pink noise with f^-1 PSD
• Brownian or random walk noise with f^-2 PSD

see /tests for tests that compare allantools output to other (e.g. Stable32) programs. More test data, benchmarks, ipython notebooks, and comparisons to known-good algorithms are welcome!

See /docs for documentation in sphinx format. On Ubuntu this requires the python-sphinx and python-numpydoc packages. html/pdf documentation using sphinx can be built locally with:

/docs$ make html
/docs$ make latexpdf

this generates html documentation in docs/_build/html and pdf documentation in docs/_build/latex.

The sphinx documentation is also auto-generated online

• http://allantools.readthedocs.org

See /examples for some examples in IPython notebook format.

github formats the notebooks into nice web-pages, for example

• https://github.com/aewallin/allantools/blob/master/examples/noise-color-demo.ipynb
• https://github.com/aewallin/allantools/blob/master/examples/three-cornered-hat-demo.ipynb

todo: add here a very short guide on how to get started with ipython

• Anders E.E. Wallin, anders.e.e.wallin "at" gmail.com
• Danny Price, https://github.com/telegraphic
• Cantwell G. Carson, carsonc "at" gmail.com
• Frédéric Meynadier, https://github.com/fmeynadier

clone from github, then install with:

sudo python setup.py install    

(see python setup.py --help install for install options)

or download from pypi:

sudo pip install allantools

New in 2016.11 : simple top-level API, using dedicated classes for data handling and plotting.

import allantools # https://github.com/aewallin/allantools/
import numpy as np

# Compute a deviation using the Dataset class
a = allantools.Dataset(data=np.random.rand(1000))
a.compute("mdev")

# Plot it using the Plot class
b = allantools.Plot()
b.plot(a, errorbars=True, grid=True)
# You can override defaults before "show" if needed
b.ax.set_xlabel("Tau (s)")
b.show()

Lower-level access to the algorithms is still possible :

import allantools # https://github.com/aewallin/allantools/ 
rate = 1/float(data_interval_in_s) # data rate in Hz 
taus = [1,2,4,8,16] #  tau-values in seconds
# fractional frequency data
(taus_used, adev, adeverror, adev_n) = allantools.adev(fract_freqdata, data_type='freq', rate=rate, taus=taus)
# phase data
(taus_used, adev, adeverror, adev_n) = allantools.adev(phasedata, data_type='phase', rate=rate, taus=taus)

# notes:
#  - taus_used may differ from taus, if taus has a non-integer multiples 
#  of data_interval - adeverror assumes 1/sqrt(adev_n) errors

The tests compare the output of allantools to other programs such as Stable32. Tests may be run using py.test (http://pytest.org). Slow tests are marked 'slow' and tests failing because of a known reason are marked 'fails'. To run all tests:

$ py.test

To exclude known failing tests:

$ py.test -m "not fails" --durations=10

To exclude tests that run slowly:

$ py.test -m "not slow" --durations=10

To exclude both (note option change):

$ py.test -k "not (slow or fails)" --durations=10

To run the above command without installing the package:

$ python setup.py test --addopts "-k 'not (fails or slow)'"

Test coverage may be obtained with the (https://pypi.python.org/pypi/coverage) module:

coverage run --source allantools setup.py test --addopts "-k 'not (fails or slow)'"
coverage report # Reports on standard output 
coverage html # Writes annotated source code as html in ./htmlcov/

On Ubuntu this requires packages python-pytest and python-coverage.

Testing on multiple python versions can be done with tox (https://testrun.org/tox)

$ tox

Tests run continuously on Travis-CI at https://travis-ci.org/aewallin/allantools