jbdatascience/gda-public

The open-source toolbox of easy-to-use topological data analysis tools, developed by GDA.

README for GDA-Public

See the thorough documentation at https://geomdata.github.io/gda-public/

This contains several fundamental tools by Geometric Data Analytics Inc. [http://www.geomdata.com] See LICENSE for copyright information.

The code is written in Python and Cython. It is written for Python 3.5, but it should work in Python 2.7 (albeit with some performance penalty). For stability and consistency, Anaconda is preferred.

See requirements.txt for requirements.

If you want to work in an air-gapped system, you can pre-download all of the required packages from http://repo.continuum.io/pkgs/ This is automated in download.py and download-viz.py, and is described below.

Please submit any BUGS to https://github.com/geomdata/gda-public/issues

Quick Start

Here are minimal instructions with no detail. First, unpack or clone the code to /path/to/gda-public. Then

bash$ cd ~   # Change directory to anywhere *except* /path/to/gda-public !
bash$ conda create --name gda_env --file /path/to/gda-public/requirements.txt python=3
bash$ source activate gda_env
(gda_env) bash$ pip install file:///path/to/gda-public
(gda_env) bash$ jupyter notebook --notebook-dir /path/to/gda-public/examples

In a worksheet, try

>>> import multidim, homology, timeseries

Building Documentation

The package comes with thorough documentation in docstrings, accessible from within python via the help( ) command:

(gda_env) bash$ python
>>> import timeseries
>>> help(timeseries)
>>> s = timeseries.Signal([1,2,3,4,5])
>>> help(s)

You can build a nice HTML guide, but you need to get a copy of the SciPy Sphinx theme:

(gda_env) bash$ cd /path/to/gda-public
(gda_env) bash$ git clone https://github.com/scipy/scipy-sphinx-theme
(gda_env) bash$ cd doc_src
(gda_env) bash$ ln -sf ../scipy-sphinx-theme/_theme ./
(gda_env) bash$ cd -
(gda_env) bash$ python setup.py build_doc_html

which can be viewed in a web browser at file:////path/to/gda-public/docs/build/html/index.html

You can also produce a static PDF documentation with

(gda_env) bash$ python setup.py build_doc_latex
(gda_env) bash$ cd /path/to/gda-public/doc_build/latex/latex
(gda_env) bash$ make

Or, just build the .tex file using your favorite TeX suite.

Running and Using

This will be filled in later versions, with suggestions.

Take a look at examples_README and the jupyter notebooks in examples/

The docstrings, accessible via help( ), are always a good reference for low-level operations.

Note! The code contains a lot of "assert" statements, so you can speed it up by using python -O.