This is a technical guide for people already familiar with scientific Python software development. For an introduction to scientific Python software development, see Become a contributor.
Atomap has a large number of unit tests, which are tested using
$ python3 -m pytest --doctest-modules atomap/
This also runs every docstring example as a unit test.
The documentation is tested by doing:
$ python3 -m pytest --doctest-glob="*.rst" doc/
Both the unit tests and the doc tests can be accelerated by running the tests in parallel. Use the xdist pytest package for this. To run the tests using 5 parallel processes:
$ python3 -m pytest -n 5 --doctest-modules atomap/ $ python3 -m pytest -n 5 --doctest-glob="*.rst" doc/
The Jupyter notebooks in https://gitlab.com/atomap/atomap_demos is also tested using pytest
$ python3 -m pytest --nbval-lax introduction_to_atomap.ipynb
Note: for some reason the
%matplotlib nbagg or
%matplotlib qt causes the tests to fail,
the easiest way of avoiding this is skipping that specific notebook cell. This done
nbval-skip to the tag for that cell.
In Atomap the PEP8 style guide is followed, and the Black code formatter is used. To automatically format the code:
$ python3 -m black atomap/
To only check the code, without changing it:
$ python3 -m black atomap/ --diff --check
Generating the sphinx page¶
These documentation pages are written by using sphinx. You generate the html site by:
$ cd doc $ python3 -m sphinx -b html ./ _build/html/
The Continuous integration (CI) settings is contained in
This runs all the above-mentioned tests, style checks and sphinx page generation on each branch.