thucyd 0.1.2

Scientific library for realtime signal processing and eigenanalysis of evolving systems.

Eigenanalysis and Filters for Signal Processing

What is thucyd?

thucyd (thoo'-sid) is an open-source library of Python-centric code that delivers implementations of eigenanalysis and causal filters that are not currently found elsewhere.

The first subpackage is eigen, and within this package a reference implementation of the algorithm for a consistent basis for eigenanalysis is provided. This subpackage is ready.

Additional subpackages will include filter_reference and filters, with an expected rolling delivery through 2020.

Package Installation

The two package hosts for thucyd are PyPi and Conda-Forge. The packages are identical and the only difference is the means of delivery. From PyPi, use pip,

$ pip install thucyd

and from Conda-Forge use conda:

$ conda install -c conda-forge thucyd

Once installed, the package is importable to Python:

>>> import thucyd

Note: At this time the conda-forge package is not yet available.

A quick example call to the eigen subpackage would be

>>> import numpy as np
>>> Vor, Eor, signs, _, _ = thucyd.eigen.orient_eigenvectors(np.eye(3).dot(np.diag([1., -1., 1.])), np.diag(np.arange(3)[::-1]))
>>> Vor
array([[1., 0., 0.],
       [0., 1., 0.],
       [0., 0., 1.]])
>>> signs
array([ 1., -1.,  1.])

Package Dependency

The only dependencies thucyd has at this time is on python >= 3.7 and numpy >= 1.14.

Why thucyd

Thucydides was the first Western writer and historian who applied scientific principles to the recording of Western history. Although Herodotus, who predates Thucydides by less than a generation, started the transformation away from the epic poetry enshrined by Homer to a more objective record, it was Thucydides who engaged in inquiry and cross validation of all accounts in his History of the Peloponnesian Wars.

The thucyd package honors the great historian by delivering implementations of eigenanalysis and signal-processing analytics that have been thoroughly researched and validated, and continues the tradition of inquiry by focusing on all the ways that rigorous eigen- and signal-processing theories can be applied to the financial markets and other machine-learning disciplines.

Buell Lane Press

Buell Lane Press is the package sponsor.