maestral/CONTRIBUTING.md

89 lines
3.6 KiB
Markdown

### Code
To start, install maestral with the `dev` extra to get all dependencies required for
development:
```
pip3 install maestral[dev]
```
This will install packages to check and enforce the code style, use pre-commit hooks and
bump the current version.
Code is formatted with [black](https://github.com/psf/black).
Coding style is checked with [flake8](http://flake8.pycqa.org).
Type hints, [PEP484](https://www.python.org/dev/peps/pep-0484/), are checked with
[mypy](http://mypy-lang.org/).
You can check the format, coding style, and type hints at the same time by running the
provided pre-commit hook from the git directory:
```bash
pre-commit run -a
```
You can also install the provided pre-commit hook to run checks on every commit. This
will however significantly slow down commits. An introduction to pre-commit commit hooks
is given at [https://pre-commit.com](https://pre-commit.com).
### Documentation
The documentation is built using [sphinx](https://www.sphinx-doc.org/en/master/) and a
few of its extensions. It is built from the develop and master branches and hosted on
[Read The Docs](https://maestral.readthedocs.io/en/latest/). If you want to build the
documentation locally, install maestral with the `docs` extra to get all required
dependencies:
```
pip3 install maestral[docs]
```
The API documentation is mostly based on doc strings. Inline comments should be used
whenever code may be difficult to understand for others.
### Tests
The test suite uses [pytest](https://pytest-cov.readthedocs.io/en/latest/) to define
and run all tests.
Test are grouped into those which require a linked Dropbox account ("linked") and those
who can run by themselves ("offline"). The former tend to be integration test while the
latter are mostly unit tests. The focus currently lies on integration tests, especially
for the sync engine, as they are easier to maintain when the implementation and internal
APIs change. Exceptions are made for performance tests, for instance for indexing and
cleaning up sync events, and for particularly complex functions that are prone to
regressions.
**Important**: The entire content of the Dropbox folder is deleted before and after
running each test case. You should therefore use a dedicated account to run tests.
The current test suite uses a Dropbox access token provided by the environment variable
`DROPBOX_ACCESS_TOKEN` or a refresh token provided by `DROPBOX_REFRESH_TOKEN` to connect
to a real account. The GitHub action which is running the tests will set the
`DROPBOX_ACCESS_TOKEN` environment variable for you with a temporary access token that
expires after 4 hours. Tests are run on `ubuntu-latest` and `macos-latest` in parallel
on different accounts.
When you create a new test case, you should acquire a "lock" on the account before
running tests to prevent them from interfering which each other by creating a folder
`test.lock` in the root of the Dropbox folder. This folder should have a
`client_modified` time set in the future, to the expiry time of the lock. Fixtures to
create and clean up a test config and to acquire a lock are provided in the
`tests/linked/conftest.py`.
If you run the tests locally, you will need to provide a refresh or access token for
your own Dropbox account. If your account is already linked with Maestral, it will have
saved a long-lived "refresh token" in your system keyring. You can access it manually or
through the Python API:
```Python
from maestral.main import Maestral
m = Maestral()
print(m.cred_storage.token)
```
You can then store the retrieved refresh token in the environment variable
`DROPBOX_REFRESH_TOKEN` to be automatically picked up by the tests.