Skip to content

Latest commit

 

History

History
110 lines (78 loc) · 4.62 KB

README.md

File metadata and controls

110 lines (78 loc) · 4.62 KB

NEURON Documentation

Online

Latest NEURON documentation is available on:

Contents:

  • Documentation on building NEURON
  • User documentation (HOC, Python, tutorials, rxd)
  • Developer documentation (SCM, technical topics, Doxygen)

Local build

Virtual environment

It is recommended to use a Python virtual environment, for example:

python3 -m venv venv
source venv/bin/activate

In order to build documentation locally, you need to pip install the docs_requirements.txt :

pip3 install -r docs/docs_requirements.txt --upgrade

Also, make sure to have Doxygen and pandoc installed, and the dependencies listed in conda_environment.yml Note that this conda environment file is tailored to the online ReadTheDocs setup (but it lists out all desired requirements, so make sure to check it out).

Anaconda environment

After installing Anaconda, create a new environment with the following command:

conda env create --quiet --name rtd --file docs/conda_environment.yml
conda activate rtd

This will install all dependencies needed to build the documentation locally, in a similar way as on ReadTheDocs. However ReadTheDocs has a different setup, so it is of interest to head over and check the build logs for additional information.

Confguring the build

With all dependencies installed, configure project with CMake (>= v3.17) as described in CMake Build Options.

e.g. in your CMake build folder:

cmake .. -DNRN_ENABLE_DOCS=ON

Building the documentation

Note that executing the jupyter notebooks requires working NEURON installation. So make sure the neuron python module can be imported. For this, you can build and install neuron from source first or you could also do a pip install neuron (in case you don't want/need to compile NEURON from source).

In order to build the full documentation (Doxygen, Jupyter notebooks, Sphinx):

make docs

That will build everything in the nrn/docs/_build folder from where you can open index.html locally.

In case you just want the Sphinx build to be performed(i.e. you are not working on Jupyter notebooks or doxygen):

make sphinx

Faster Local Iterations

When working locally on documentation, depending on what you work on, be aware of the following targets to speed up building process:

  • doxygen - build the API documentation only. Ends up in _generated folder under docs.
  • notebooks - execute & embed outputs in-place into jupyter notebooks, see notebooks.sh
  • notebooks-clean - clears outputs from notebooks. Remember that executing notebooks will add outputs in-place, and we don't want those committed to the repo.
  • sphinx - build Sphinx documentation

NOTE:

  • docs target calls: doxygen notebooks sphinx and notebooks-clean.
  • sphinx target is the one that will assemble all generated output (doxygen, notebooks).

ReadTheDocs setup

Config file

Configuration file is in the top directory: ../.readthedocs.yml.

Documentation not handled by RTD

For RTD we need to call doxygen and notebooks make targets, since only sphinx executable is called therein. To that end we are using conda to be able to install packages like CMake in RTD. All conda dependencies are listed under conda_environment.yml.

Leveraging conf.py

The only way to gather extra documentation in RTD is to make use of conf.py. Have a look at if os.environ.get("READTHEDOCS"): block in conf.py to see how we leverage the aforementioned targets and outputs.

Notebooks execution with neuron wheels

For notebooks we are using neuron Python wheel as follows:

  • neuron-nightly for master RTD builds.
  • neuron==X.Y.Z for release RTD builds. Note that wheels need to be published first (see next section).

To achieve this, we parse the RTD environment variable READTHEDOCS_VERSION and pip install it (see conf.py) .

New release on RTD

Follow the following steps:

NOTE: build can be re-launched as many times as needed