github.com/weihuayi/fealpy

We want to develop an efficient and easy to use finite element software package to support our teach and research work.

We still have lot work to do.

In fealpy/, run the following command

sudo pip install .

whill copy the fealpy into /usr/local/lib/python3.x/dist-packages/

In the fealpy/, run the following command:

sudo pip install -e .

which will create a link in /usr/local/lib/python3.x/dist-packages/.

sudo apt-get install python3 python3-pip ipython3 
sudo apt-get install python3-numpy
sudo apt-get install python3-scipy
sudo -H pip3 install matplotlib
sudo -H pip3 install plotly

sudo apt-get install petsc-dev
sudo -H pip3 install petsc4py 

sudo apt-get install python3-skimage

Install mesh generation package meshpy:

ububtu 18.04 install meshpy

sudo apt-get install libboost-python-dev
sudo -H pip3 install meshpy

Install "metis"

sudo apt-get install metis

sudo apt-get install python3-numexpr 

The following can install by pip3

sudo -H pip3 install pyamg
sudo -H pip3 install pycuda
sudo -H pip3 install mpi4py
sudo -H pip3 install pymetis
sudo -H pip3 install PySPH
sudo -H pip3 install nipy
sudo -H pip3 install pygmsh

conda install pyamg
conda install -c conda-forge pytools

from https://www.lfd.uci.edu/~gohlke/pythonlibs/#meshpy, dowload

'MeshPy‑2018.1‑cp36‑cp36m‑win_amd64.whl'

pip install msgpack
pip install MeshPy‑2018.1‑cp36‑cp36m‑win_amd64.whl

Dowload the latest boost from http://www.boost.org/.

./bootstrap.sh --with-python=python3
./b2

Here we install boost into /usr/local directory.

We also use VTK Python 3 wrapper, one can clone VTK from github first:

git clone https://github.com/Kitware/VTK.git
cd VTK

Then check out the latest stable version:

git archive v7.1.0 --prefix=vtk7.1.0/ --format=zip  > ../vtk7.1.0.zip

Unzip the vtk7.1.0.zip, then cd the source code file:

cd ..
unzip vtk7.1.0.zip
cd vtk7.1.0/

Create a build directory, then cd into it and run cmake-gui ..

mkdir build
cd build/
cmake-gui ..

You can see the GUI of cmake, then:

1. check the Advanced option
2. Press Configure button, and a dialog box will pop up
3. Press the Finish button on the dialog box
4. When configure completed, input test in search input box, uncheck BUILD_TESTING. If you check this argument, vtk will dowload many data which is very slow.
5. Then, replace keyword test by python in search input box
   • Set PYTHON_EXECUTABLE as /usr/bin/python3.5
   • Set VTK_PYTHON_VERSION as 3.5
   • Check VTK_WRAP_PYTHON
6. Press Configure button again
7. When configure completed, press Generate button
8. When generate process completed, just close the GUI of cmake

Finally, we can compile VTK by make command

make -j8 # parallel compile with 8 thread (My laptop have 8 core)
sudo make install # it will intsall vtk into /usr/local/

Then one should append the following lines into your bashrc file:

# if you install vtk into /usr/local
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib
export PYTHONPATH=$PYTHONPATH:/usr/local/lib/python3.5/site-packages

# if you install vtk into /home/why/software/vtk/7.1.0/
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/why/software/vtk/7.1.0/lib
export PYTHONPATH=$PYTHONPATH:/home/why/software/vtk/7.1.0/lib/python3.5/site-packages

After you update your bashrc file, please source it:

source ~/.bashrc

Then you can use vtk in your python 3 program or ipython3, just:

import vtk

At last, one can intall mayavi, but one should change into root user first:

$ sudo -s
# export PYTHONPATH=$PYTHONPATH:/home/why/software/vtk/7.1.0/lib/python3.5/site-packages
# pip3 install mayavi

sudo apt-get install python3-dbg

Debug python program:

$ gdb python3
...
(gdb) run <programname>.py <arguments>

• https://plot.ly/python/
• http://www.math.uci.edu/~chenlong/programming.html