neXtSIM_DG : next generation sea-ice model with DG

Here you will find the sources for neXtSIM_DG sea-ice model.

Installation instructions can be found here.

Below are some conventions that every contributors to this model must apply in order to have the most efficient and consistent workflow for the model developments. These conventions deal with :

• coding
• commenting
• code management

Coding conventions

For neXtSIM_DG we use clang-format and the Webkit style, with a 100 character line length limit.

• neXtSIM is written using ISO C++17
• All array operations should be done using std::vectors - not C-style arrays
• The use of C-style pointers, new, and delete is strongly discouraged
• Names and values of physical constants reside in src/include/constants.hpp
• Runtime modifiable model parameters are set in src/options.cpp and can be set using option files.

A dedicated clang format file has been designed for the code. You may run it locally and manually with the command clang-format -i $yourfile or have a plug-in with your favourite code editor or implement a git pre-commit hook locally by putting this pre-commit file in your .git/hooks/. This clang formatting will also be run each time a pull request is done as part of the continuous integration (see continuous integration script.

For each file file.cpp in src/, an associated test file file_test.cpp shall be created in test/ and contain all the associated tests. The executable produced by these files must start with the sub-string test. As an example the test file for the Iterator file produce testIterator executable

Commenting conventions for a nice automatic documentation

The automatic generation of documentation is done with Doxygen and Sphinx. The update of the documentation each time the code is modified is done with readthedocs and github actions.

In order to have the most understandable and consistent documentation of the API, we expect that contributors to neXtSIM write some comments alongside their code for every C++ component (namespaces, types, methods, functions, and constants) they implement.

Every file must begin with :

• a description of what it contains
• the name of the authors
• the date of creation/modification

Example :

/*!
* @file   gridoutput.hpp
* @author Einar Olason <einar.olason@nersc.no>
* @date   Thu Aug  4 09:47:27 CEST 2016
*/

We ask every coders to follow the following commenting conventions for comments providing the automatic documentation:

• documentation blocks must appear before the component it describes, with the same indentation (classes and functions),
• comments that should appear in the automatic documentation start with /*! or //!
• multi-line documentation block must begin with /*! and end with */ on their own line,
• single-line block begins with //!,
• description of variables or constants on the same line they are declared must preceded by //!<,
• documentation must use tags like @see or @param and Markdown for formatting.

Comments starting with // or /* won’t appear in the automatic documentation so they can be used for internal or development comments.

The structure of a documentation block for classes and functions should be organized as follow :

• short summary
• extended summary (optional)
• template parameters (for classes, methods and functions)
• function/method parameters (for methods and functions)
• returns (for methods and functions)
• throws (for methods and functions)
• exception safety (for methods and functions)
• helper functions (for functions)
• initializer declaration (for constants)
• see also (optional)
• notes (optional)
• references (optional)
• examples (optional)

Examples :

/*!
 * Sum numbers in a vector.
 *
 * @param values Container whose values are summed.
 * @return sum of `values`, or 0.0 if `values` is empty.
 */
double sum(std::vector<double> & const values) {
    ...
}

/*!
 * Compute an element-wise cosine.
 *
 * @see sin
 * @see tan
 * @see [numpy.vectorize](https://docs.scipy.org/doc/numpy/reference/generated/numpy.vectorize.html)
 */
vector<double> cos(vector<double> const & angles);

/*!
 * This is an amazing function! For example:
 *
 *     auto cosines = cos(angles);
 *
 * Comment explaining the second example.
 *
 *     auto cosines = cos(radians(angles));
 */

/*!
 * Supported coordinate systems for flux-conserving transformations.
 *
 * These values are used in arguments to various functions in this package.
 * Unless otherwise stated, these functions do not validate whether the data
 * set makes sense in the "from" coordinates.
 */
enum class CoordType
{
    PIXEL,    //!< Untransformed detector coordinates
    WARP_PIXEL,    //!< Distortion-corrected detector coordinates
    SKY_WCS    //!< Equatorial J2000.0 coordinates
};

/*!
 * Read an image from disk.
 *
 * @param fileName the file to read. Must be either absolute or relative to
 *     the program working directory.
 *
 * @return the image stored in `fileName`. If the image on disk does not
 *     have `double` pixels, they will be cast to `double`.
 *
 * @throws IoError Thrown if `fileName` does not exist or is not readable.
 *
 * @exceptsafe Strong exception guarantee.
 */
lsst::afw::image::Image<double> loadImage(std::string const & fileName);

The comments providing automatic documentation should be supplemented with ordinary explanatory and internal comments using the standard comment syntax of /* and //. These will not be picked up by the automatic documentation and can take a more general format.

Code management on github

Version numbering

We use a derived version of semantic versioning. In brief this means the main branch has a version number assigned (tagged) to each commit. The numbers are of the form x.y.z, where we:

1. increment z when there's a hotfix (bugfix) pushed to main
2. increment y when develop is merged into main
3. increment x when we've been working on something extraordinary, when there's a significant change in an interface (configs or outputs) ... or when y is becoming too big.

Tag and releases

To implement the version numbering two steps are required: first, you have to tag a version of the code with the appropriate name (v1.0.0 for instance the rules described above) and then a release can be generated from the tag (same name).

Both operations need to be commented, with git tag -a v1.0.0 -m 'description of the tagged version of the code' when generating a tag and in the Release title and description fields on github when generating the release so that changelog files are informative for users.

After each release, the changelog file must be generated as described here and deployed in the docs branch. This changelog file can also be modified anytime to take into account the recent developments.

Git branching and merging

For the code we adopt the following system for branching and merging:

1. main branch: numbered releases of the code. Never edited. Merged from develop and hot fix branches (see notes on workflow below). Long living.
2. develop branch: rather stable version of code under development. Never edited. Merged from topic specific issue branches. Long living.
3. feature<NNN>_<short_heading> : feature branches to develop significant components. Related to one or more issues, which should be linked to the relevant draft pull request (NNN = draft pull request number). Long living. Branched from, regularly updated from and merged back into develop.
4. issue<NNN>_<short-heading> : issue specific branches (NNN = issue number). Main working area. Short living. Branched from, and merged back into develop or a feature branch.
5. hotfix<NNN>_<short-heading> : branches that are specific to a hotfix issue. Hotfixes are bugfixes on main that should be fixed as soon as possible.

Note :

1. Never edit code in the main or develop branch. Always make a new branch for your edits.
2. A new branch should be very specific to only one problem. It should be short lived.
3. Commit often.
4. Branch often.
5. Branch only from main (hotfixes), develop (features and issues) or feature branches (feature related-issues).
6. Create pull requests for your branches and always assign a reviewer to merge and delete the branch, and close the issue. Commit messages should be formatted as a short description, followed by an empty line, followed by a detailed description. Line wrapping at 80 characters is preferred.
7. Every commit must compile.
   • With one exception: If it is necessary to commit code that does not compile, for example as part of a large refactor, then the commit message must be prefixed by WIP (for Work In Progress). This will make it clear to other developers that it is a commit that does not compile.

How to report and handle new issues (bugs, improvements, new features, etc.)

If you discover a bug check that no one else has reported the same issue on https://github.com/nextsimdg/nextsimdg/issues. If the bug has not been reported, create an issue and assign someone to fix it (possibly yourself). Please notify people in person if you're assigning issues to them.

If you would like to suggest an improvement or a new feature check that no one else has made a similar request on https://github.com/nextsimdg/nextsimdg/issues. If this is not the case create a new issue, assigning or mentioning anyone you think could be affected or interested by your suggestion.

If you have been assigned an issue on https://github.com/nextsimdg/nextsimdf/issues address it using the following steps. For issues requiring a hotfix (bugs on the main branch that should be fixed as soon as possible):

1. Branching off from main, create an issue branch on your local system named hotfix<NNN>_<short-heading> where NNN is the issue number from GitHub. This will be the main (short living) working area.
2. Write the necessary code.
   • a. Make sure you test your modifications well.
   • b. Feel free to commit and push your issue branch often
3. Once the issue is fixed merge the main branch back into your issue branch and resolve any conflicts.
4. Create a pull request on GitHub to merge the issue branch back into main, always include at least one reviewer who will then merge and delete the issue branch.
5. Merge the main branch into develop in your local repository, resolve conflicts, test, and push the updated develop branch.
6. Tag the merge by incrementing the patch number of the version number.
   • a. Do this on the command line with git tag LABEL.
   • b. See the existing tags with git tag -l.
7. Close the issue.

For issues not requiring a hotfix (less urgent bug-fixes and feature requests):

1. Branching off from develop, create an issue branch on your local system named issue<NNN>_<short-heading> where NNN is the issue number from GitHub. This will be the main (short living) working area.
2. Write the necessary code.
   • a. Make sure you test your modifications well.
   • b. Feel free to commit and push your issue branch often
3. Once the issue is fixed merge the develop branch back into your issue branch and resolve any conflicts.
4. Create a pull request on GitHub to merge the issue branch back into develop. Always include at least one reviewer who will then merge and delete the issue branch, and close the issue.

doctest

Nextsim-DG uses the doctest library for testing. The doctest library is copyright Viktor Kirilov and written by Viktor Kirilov and the other doctest contributors. The library is licensed under the MIT license, which can be found in the lib/doctest subdirectory.