ssalinasfe / polylla-mesh

New simplest version using POO and halfedge data structure

There is a new version of this mesh generator to facilitate the integration with third party software.

https://github.com/ssalinasfe/Polylla-Mesh-DCEL

POLYLLA: Polygonal meshing algorithm based on terminal-edge regions

New algorithm to generate polygonal meshes of arbitrary shape, using any kind of triangulation as input, adaptable to any kind of geometry, no addition of extra points and with a data structure tp an easy implementation to all language programmings.

The algorithm needs a triangulation as input, this triangulation is represent using 3 files:

• .node: Point file
• .ele: Triangle file
• .neigh: adjacency list

Example of use:

Calls to file with points:

./Polylla <node file> <ele file> <neigh file> <output filename>

The output is a <output filename>.off file, a <output filename>.svg file and a <output filename>.json file with the metrics of the mesh.

Input data generation

Following scripts to generate input files were made to test the algorithm using the Virtual Element Method

Random points files generation (.node)

• 2x2RandomPoints.py: Generate random points inside a square 2x2

Random point PLSG (.poly)

• R2DiskRandom.py: Generate a PLSG of a circle of radio 2 with random points inside of it
• 2x2Lrandom.py: Generate a PLSG with L-shape with random points inside of it
• 2x2QuaterCircle.py: Generate a PLSG in a square of 2x2 with a quartercircle of radio 0.4 in the left botton corner and random points inside of it

Semiuniform triangulations (.node and .ele)

• R2DiskSemiuniform.py: Generate a semiuniform triangulation in a disk of radio 2
• 2x2QuaterCircleUniform.py: Generate a semi Uniform triangulation inside a square of 2x2 with a quartercircle of radio 0.4 in the left botton corner.
• Luniform: Generate a semi uniform triangulation with L-shape

Work in progress. TODO

• Agregar que acepte cualquier input mientras que tengan un archivo .node. .ele, -neigh
• Cambiar stream de archivo por stream de string en la funcion read_from_triangle
• Verificar la estructura de datos del input para que haya una correcta asociación de los triangulos y la adjacencia
• Aceptar como input .node and .ele y generar la adjacencia en base a eso.
• Usar safe arrays en vez de array
• Cambiar polygon array por list o vector
• Eliminar funciones no usadas
• Reescribir todo y encapsularlo en una clase para usarlo como librería
• Llamar a triangle en vez de a detri2 cuadno solo se tenga de input un .node (detri no genera triangulaciones de más 10^6 puntos)
• Generar como output un svg
• Construir test