vc-flas 0.1.1

Python implementation of the Fast Linear Assigment Sorting Algorithm

Fast Linear Assignment Sorting - FLAS

FLAS allows to sort 2d grids based on similarity. This directory contains the python implementation and shows how to use it.

Features

• Implementation of FLAS and DPQ (and other metrics to quantify an arrangement)
• Allows frozen fields
• Allows empty fields (holes)
• Deterministic random when seeded
• Installable via PyPI
• Easy to use python interface with numpy support
• Fast implementation in C++

Installation

pip install vc_flas

Install from sources

git clone https://github.com/Visual-Computing/LAS_FLAS.git
cd LAS_FLAS/c_python
pip install .  # or
pip install -e .  # for develop/editable mode

Usage

Basic Usage

Given that you have N features with D-dimensions, you can sort these features based on similarity with the following code:

import numpy as np
from vc_flas import Grid, flas

N, D = 241, 3
features = np.random.random((N, D))
grid = Grid.from_features(features)

arrangement = flas(grid, wrap=True, radius_decay=0.99)

sorted_features = arrangement.get_sorted_features()
height, width, dim = sorted_features.shape
assert (height, width, dim) == (16, 16, 3)

# show_image(sorted_features)

Working with Arrangements and Labels

Often you need not only the features sorted, but other objects (like images for example) as well. See this example for more information on that.

Creating Grids

There are more ways to initialize grids. See here for some examples.

Metrics

TODO

About

Kai Barthel, Nico Hezel, Klaus Jung, Bruno Schilling and Konstantin Schall

HTW Berlin, Visual Computing Group, Germany

Visual Computing Group

This is an example implementation of the algorithms from the paper

Improved Evaluation and Generation of Grid Layouts using Distance Preservation Quality and Linear Assignment Sorting

Published in COMPUTER GRAPHICS Forum: (https://onlinelibrary.wiley.com/doi/full/10.1111/cgf.14718)

Reference

Reference to cite when you use any of the presented algorithms in a research paper:

@article{https://doi.org/10.1111/cgf.14718,
    author = {Barthel, K. U. and Hezel, N. and Jung, K. and Schall, K.},
    title = {Improved Evaluation and Generation Of Grid Layouts Using Distance Preservation Quality and Linear Assignment Sorting},
    journal = {Computer Graphics Forum},
    volume = {42},
    number = {1},
    pages = {261-276},
    keywords = {interaction, user studies, visualization, information visualization, high dimensional sorting, assistive interfaces},
    doi = {https://doi.org/10.1111/cgf.14718},
    url = {https://onlinelibrary.wiley.com/doi/abs/10.1111/cgf.14718},
    year = {2023}
}

How to publish a new version

• Run git checkout main and git pull to be sure, all updates are fetched
• Edit version number in c_python/src/vc_flas/__init__.py to x.y.z
• Run git add -A, git commit -m 'vx.y.z' and git tag -a vx.y.z -m 'vx.y.z'
• Run git push and git push origin --tags