Skip to content

Latest commit

 

History

History
206 lines (142 loc) · 5.99 KB

README.rst

File metadata and controls

206 lines (142 loc) · 5.99 KB
Raw

CAID

CAID is a multi-platform software that has been designed for IsoGeometric Analysis Pre and Post Processing. Its design goal is to provide a fast, light and user-friendly designer and meshing tool.

The Post-Processing and advanced visualization capabilities are still under development, as well as an optimized 3D computing engine.

For more details, please read [CAID](http://ratnani.org/caid_doc/)

CAID can be used in 2 different ways:

  • as a python package: using ipython or scripts
  • as a GUI modeler: for a better interactive use

In order to follow related developments to CAID, please visit our JOREK-Django project on (https://tree.taiga.io/project/ratnani-jorek-django/)

Common requirements

  • Python3 interpreter (https://docs.python.org/3):

    sudo apt-get install python3 python3-dev

  • pip package manager for Python (https://pypi.python.org/pypi/pip):

    sudo apt-get install python3-pip

  • Python dependencies (Numpy, Scipy, Matplotlib) can be installed using:

    python3 -m pip install --user -r requirements.txt

  • igakit (http://bitbucket.org/dalcinl/igakit) is a Python package that implements many of the NURBS routines from Piegl's book using Fortran and Python:

    python3 -m pip install --user hg+https://bitbucket.org/dalcinl/igakit

Additional GUI requirements

  • wxPython (https://wxpython.org), a cross-platform GUI toolkit for the Python language.

    To install on Windows and Mac OS X:

    python3 -m pip install --user wxPython

    To install on Linux, correct binaries must be selected depending on distribution (check availability at https://extras.wxpython.org/wxPython4/extras/linux/gtk3). For example:

    python3 -m pip install --user -f https://extras.wxpython.org/wxPython4/extras/linux/gtk3/ubuntu-16.04 wxPython

  • PyOpenGL (http://pyopengl.sourceforge.net/), the most common cross platform Python binding to OpenGL and related APIs:

    python3 -m pip install --user PyOpenGL PyOpenGL_accelerate

Installing CAID

  • Standard mode:

    python3 -m pip install --user .

  • Developer mode:

    python3 -m pip install --user -e .

Package-Usage

Start by import the CAID package:

import caid

In order to check that the fortran code has been well installed, use:

import caid.core
import caid.core.bspline
import caid.core.hbezier

If you don't get any message, then everything seems to be fine.

Let's now import a square geometry and refine it:

from caid.cad_geometry import square
geo = square(n=[3,3], p=[2,2])
geo.plotMesh()
import matplotlib.pyplot as plt
plt.show()

A cad_geometry object contains a list of cad_nurbs (more generally any class that inherits from the cad_object class) with some additional informations. The class cad_nurbs inherits the NURBS class in igakit. It represents a single patch but with some additional informations like

  • orientation: needed for Neumann boundary conditions
  • rational: True if we use the weights. Default value : False

In the future, the cad_geometry class will also contain Splines on triangulations (using splitri)

GUI-Usage

Add the following lines in your .bashrc/.bash_profile by replacing PATH_TO_CAID_SRC with the correct path to the project source directory:

alias caid="python3 $PATH_TO_CAID_SRC/caid-gui/main.py"

Running CAID can be done in different ways.

  • without any argument

    caid

  • with a given workgroup session

    caid session.wkl

  • with given domains files session

    caid domain1.xml domain2.xml domain3.xml

  • with a given field file

    caid U.pfl

TODO

  • update setup file, using pip to install all dependencies.