Skip to content

Latest commit

 

History

History
127 lines (87 loc) · 6.35 KB

README.md

File metadata and controls

127 lines (87 loc) · 6.35 KB
Raw

Nbviewer Binder Build Status

Coverage Status

Index of jupyter notebooks

About

This repository contains notebook-based documentation for AbiPy. This augments our Sphinx-based documentation with jupyter notebooks containing interactive tutorials and examples. Additional examples are available on the:

How to use the tutorials

The repository contains the input required to run the lessons on Abipy, as well as the input files, the main output in text format and the netcdf files. The workhorse of these tutorials are so-called 'jupyter notebooks'. There are several options available to you depending on the software installed on your machine.

You can:

  1. Follow the tutorial using the static HTML version, and look at the input/output files in jupyter notebooks in the github repository without installing AbiPy and Abinit.

  2. Click the Launch Binder badge to start a Docker image. The image contains Abinit, AbiPy and all the other python dependencies required to run the code inside the jupyter notebooks. The notebook will be opened in your browser after building. Go into the abitutorials directory and click the index.ipynb file to open the index file. Select e.g. the Structure notebook and start to run the python code in the jupyter cells (select the cell and click the Run button). See also the other options available in the Cell tab.

  3. Install AbiPy, Abinit and Jupyter on your machine (see later) and use the python scripts as well as the netcdf files in the github repository (you also need git), like a real pythonista. You will need to execute the jupyter notebooks, and thus install all the required dependencies: python, jupyter, abipy, abinit and obviously a web browser (DOH!).

Of course, choosing between these options depends on what is your actual interest with Abipy. You might only be interested in getting a flavour of how to use Abipy (or what are these Abipy tutorials), without actually using Abipy. Then options 1 or 2 are convenient.

However, if you really plan to use Abipy, we suggest you choose 2 or 3. Really running the examples is the most efficient use of the tutorial. Of course, in order to use Abipy, you will obviously need to install it, as well as to install abinit. Choosing 2 only spare you the possible trouble of installing jupyter. The installation of Abipy and abinit (also using git) in a coherent way is presented in the Abipy README on Github, or in the stable version of the Abipy doc,
or in the develop version of the Abipy doc.

If you opt for option 3, after installing jupyter as described below, in addition to AbiPy and abinit, you will have to open the notebook in your browser

jupyter notebook FILE.ipynb

where FILE.ipynb is one the ipython notebooks available inside the abitutorials directory. To open, for instance, the notebook for the first lesson, use:

cd abitutorials
jupyter notebook base1/lesson_base1.ipynb

Acccessing the notebooks and installing jupyter

First step, download the abitutorials

git clone https://github.com/abinit/abitutorials

In case you followed the conda way to install Abipy and abinit, the installation of jupyter is very simple. Be sure to install it in the same conda environment as Abipy and abinit, though.

Inside the abitutorials directory, issue

conda install --file ./requirements.txt

Then, you might also like to install some other python packages, that are used in selected tutorials or for specific abipy commands. Among the list, graphviz and python-graphviz, in order to allow the full vizualization of figures in the tutorials, and pybtex if you want to extract references from CIF files .

conda install graphviz
conda install python-graphviz 
conda install pybtex

More detailed instructions on how to install with conda are available in the abiconda documentation.

As an alternative to conda, you can use pip to install python code for jupyter with:

pip install jupyter abipy

Note, however, that you may need to compile from source some of the low-level dependencies including the Abinit executable so the entire process could require some time, besides your machine must provide a sane environment to build Fortran/C/C++ code, possibly with MPI support.

If you love building software from source, feel free to use this approach and use the configuration examples available on the abiconfig repository to build Abinit on your machine/cluster.

After installing jupyter, we can launch one of the notebooks as described above.

As a test of the coherent installation of AbiPy and abinit, select one of the directory with a lesson_*.py script e.g. sigeph/lesson_sigeph.py. Read the corresponding README.md file. Then look at the python script and use:

./lesson_sigeph.py

to run the flow automatically or

./lesson_sigeph.py --help

to get the list of supported options.

TIP:

All the AbiPy scripts start with the abi prefix. Just type:

abi + <TAB>

in the shell to get the list of possible scripts. Please consult the AbiPy documentation for further details.