maljovec/regulus

Regulus

Visual analytics of high-dimensional data using a topology based approach. Regulus support both Morse and Morse-Smale complexes.

Installation

Server

Regulus uses Bottle python server, which can installed via pip install bottle or conda install bottle -c conda-forge.

Client

Install the yarn package manager. e.g. conda install yarn -c conda-forge.

• Run yarn install in the regulus root directory. It will locally install all the required Javascript packages.
• Run yarn build to build the Javascript code.
• During development, use yarn watch to continuously rebuild Regulus you save files.

Data

Regulus works with high-dimensional data where the first d-dimensions are independent variables (dimensions) and the rest of dependent variables (measures).

• todo: [expand]: convert csv file to a regulus json file
• todo: [expand]: compute Morse- or Morse-Smale complexes
• todo: [expand]: compute various statistic such as linear regression and inverser regression

Data sampling

Regulus can sample the domain to create additional data points during the analysis process. There are two main approaches,

1. Interpolation of existing data points

TBD

2. Running a user simulation program

Currently, Regulus only support running a local cyclus for a specific scenario.

Running

Server:

cd server; python server.py -d <path-to-data-directory>

You can also set an environment variable REGULUS_DATA_DIR to point at the data directory.

Client

Point your browser to localhost:8081

Regulus automatically save and reload the layout configuration.

• localhost:8081/?noload prevents loading of previous layout
• localhost:8081/?nosave prevents saving the layout during this run
• localhost:8081/?noload&nosave prevents both