Developer Manual

Shareloc is an open source software : don’t hesitate to hack it and contribute !

Please go to the GitHub repository for source code.

Read also Shareloc Contribution guide with LICENCE

You need also to complete, sign and email to cars [at] cnes [dot] fr the Individual Contributor Licence Agrement or Corporate Contributor Licence Agrement.

Contact: cars AT cnes.fr

Developer Install

Install from source procedure is followed.

The detailed development install method is described in the Makefile

Particularly, it uses the following pip editable install:

pip install -e .[dev]

With this pip install mode, source code modifications directly impacts Shareloc code.

Coding guide

Here are some rules to apply when developing a new functionality:

  • Comments: Include a comments ratio high enough and use explicit variables names. A comment by code block of several lines is necessary to explain a new functionality.

  • Test: Each new functionality shall have a corresponding test in its module’s test file. This test shall, if possible, check the function’s outputs and the corresponding degraded cases.

  • Documentation: All functions shall be documented (object, parameters, return values).

  • Use type hints: Use the type hints provided by the typing python module.

  • Use doctype: Follow sphinx default doctype for automatic API

  • Quality code: Correct project quality code errors with pre-commit automatic workflow (see below)

  • Factorization: Factorize the code as much as possible. The command line tools shall only include the main workflow and rely on the shareloc python modules.

  • Be careful with user interface upgrade: If major modifications of the user interface or of the tool’s behaviour are done, update the user documentation (and the notebooks if necessary).

  • Logging and no print: The usage of the print() function is forbidden: use the logging python standard module instead.

  • Limit classes: If possible, limit the use of classes as much as possible and opt for a functional approach. The classes are reserved for data modelling if it is impossible to do so using xarray and for the good level of modularity.

  • Limit new dependencies: Do not add new dependencies unless it is absolutely necessary, and only if it has a permissive license.

Pre-commit validation

A pre-commit validation is installed with code quality tools (see below). It is installed automatically by make install-dev command.

Here is the way to install it manually:

$ pre-commit install

This installs the pre-commit hook in .git/hooks/pre-commit from .pre-commit-config.yaml file configuration.

It is possible to test pre-commit before commiting:

$ pre-commit run --all-files                # Run all hooks on all files
$ pre-commit run --files shareloc/__init__.py   # Run all hooks on one file
$ pre-commit run pylint                     # Run only pylint hook

Code quality

Shareloc uses Black, and Pylint quality code checking.

Use the following command in shareloc virtualenv to check the code with these tools:

$ make lint

Use the following command to format the code with black:

$ make format

Black

Black is a quick and deterministic code formatter to help focus on the content.

Shareloc black configuration is done in pyproject.toml

If necessary, Black doesn’t reformat blocks that start with “# fmt: off” and end with # fmt: on, or lines that ends with “# fmt: skip”. “# fmt: on/off” have to be on the same level of indentation.

Black manual usage examples:

$ cd SHARELOC_HOME
$ black --check shareloc tests  # Check code with black with no modifications
$ black --diff shareloc tests   # Show black diff modifications
$ black shareloc tests          # Apply modifications

Pylint

Pylint is a global linting tool which helps to have many information on source code.

shareloc pylint configuration is done in dedicated .pylintrc file.

Pylint messages can be avoided (in particular cases !) adding “# pylint: disable=error-message-name” in the file or line. Look at examples in source code.

Pylint manual usage examples:

$ cd SHARELOC_HOME
$ pylint tests shareloc       # Run all pylint tests
$ pylint --list-msgs          # Get pylint detailed errors informations

Tests

Shareloc includes a set of tests executed with pytest tool.

To launch tests:

make test

Advanced testing

To execute the tests manually, use pytest at the Shareloc projects’s root (after initializing the environment):

$ python -m pytest

To run only the unit tests:

$ cd shareloc/
$ pytest -m unit_tests

It is possible to obtain the code coverage level of the tests by installing the pytest-cov module and use the --cov option.

$ cd shareloc/
$ python -m pytest --cov=shareloc

It is also possible to execute only a specific part of the test, either by indicating the test file to run:

$ cd shareloc/
$ python -m pytest tests/test_triangulation.py

Or by using the -k option which will execute the tests which names contain the option’s value:

$ cd shareloc/
$ python -m pytest -k triangulation

By default, pytest does not display the traces generated by the tests but only the tests’ status (passed or failed). To get all traces, the following options have to be added to the command line (which can be combined with the previous options):

$ cd shareloc/
$ python -m pytest -s -o log_cli=true -o log_cli_level=INFO

C++ bindings with Pybind11

This paragraph describes the use of C++ codes inside the project.

To improve execution time when creating epipolar grids, we use C++ bindings from the python code.To do this, we use the RPCoptim class, which is a copy of the RPC class but whose methods call up C++ code via bindings. As a result, the methods of RPCoptim have a shorter execution time than the methods of the RPC class. (This is not always true, as python methods use numba to speed up the execution of python code. So, in some cases python methods have a slightly shorter execution time than C++ methods).

As the C++ code is compiled during the installation of shareloc, the installation files (setup.py, setup.cfg, etc.) have been slightly modified. You are therefore strongly advised to run the ‘pip install .’ command before committing to check that the installation is working correctly. You should incorporate this command into your test routine to avoid any problems.

How does C++ bindings work ?

During installation, Pynbind11 compiles the C++ code into a “.so” file (bindings_cpp.cpython-310-x86_64-linux-gnu.so) located at the root of the project. This file can then be called in python code like any other module with the following syntax:

import bindings_cpp

bindings_cpp.RPC.__init__(self,*args,**kwargs)#for example

For this to work, we need to write a C++ code (called bind.cpp here) that tells pybind11 how to compile the C++ code so that it can be used in python. In layman’s terms, this file contains the architecture of the cpp code in the manner of a cpp project header file (.h). We should therefore remember to update it each time we modify the header file (.h). For more information, please refer to the Pybind11 documentation https://pybind11.readthedocs.io/en/stable/basics.html.

You can compile the bindings without going through the shareloc installation by using the following line command in the bind.cpp folder:

$ c++ -O3 -Wall -shared -std=c++14 -fPIC $(python3 -m pybind11 --includes) bind.cpp -o bindings_cpp$(python3-config --extension-suffix)

This is very useful when debugging.

RPCoptim Class

As explained above, RPCoptim is a copy of the Rpc class. As a result, it has exactly the same usage and its methods have the same I/O. It can be seen as a python overlay of the C++ RPC class, which acts as an interface between the python code and the C++ code.

RPCoptim inherits from RPC(C++) and GeoModelTemplate. Most of its methods consist of calling the corresponding method of the parent class in C++ and returning the output.