Skip to main content

Have your imports passed inspection?

Project description

Bonded

Do your imports pass inspection?

bonded - do your imports pass inspection?

Bonded is a linter that alerts on both missing and unused requirements.

Bonded checks for project requirements that are not actually used in the project and for imports that don't map back to any requirement explicitly declared as a dependency. By verifying both relationships, projects can be assured that all requirements necessary at runtime are properly captured as direct dependencies and not available only because of an indirect relationship. Projects can also be assured that the requirements that are declared are all necessary to the project.

Usage

Installing

pip install bonded

Running

bonded my_project_dir

By default bonded will read your pyproject.toml and find all packages or modules under the given directory. If you maintain requirements across multiple locations, you will have to tell bonded where to look.

bonded --requirements dev-requirements.txt --exclude '.*/' ./

For more examples, check out Advanced Usage.

How does it work?

Bonded searches for all imports of python modules, both explicit and implicit and associates each with an installed package. Additionally, bonded will also note the use of plugins and if the extended package is being used, the package providing the extended behavior will also be marked as used. Finally, bonded knows which packages provide executable commands that can be run on the command line and if those commands are executed, will mark the providing package as used.

If none of the above can be found for a package, it is assumed to be unnecessary to the project and is flagged so it can be removed, making refactoring requirements safer.

Bonded also remembers all imports it found while scanning for used packages, and any that were unable to be matched to installed packages are flagged as potentially missing dependencies in the package declaration.

Advanced Usage

Options

Supported command line options are:

usage: bonded [-h] [--pyproject PYPROJECT] [--setup SETUP]
              [--packages PACKAGES [PACKAGES ...]] [-r REQUIREMENTS]
              [--ignore-modules IGNORE_MODULES [IGNORE_MODULES ...]]
              [--ignore-packages IGNORE_PACKAGES [IGNORE_PACKAGES ...]]
              [--exclude EXCLUDE] [--report {table,extended-table,line,none}]
              [--verbose] [--quiet]
              [search_path]

positional arguments:
  search_path

options:
  -h, --help            show this help message and exit
  --pyproject PYPROJECT
                        Path to a pyproject.toml which will be searched for
                        requirements and bonded settings
  --setup SETUP         Path to a setup.cfg which will be searched for
                        requirements
  --packages PACKAGES [PACKAGES ...]
                        Add a package to be checked for
  -r REQUIREMENTS, --requirements REQUIREMENTS
                        Pip-requirements file used to specify further
                        requirements. Can be specified multiple times
  --ignore-modules IGNORE_MODULES [IGNORE_MODULES ...]
                        These module will not be reported as missing a package
  --ignore-packages IGNORE_PACKAGES [IGNORE_PACKAGES ...]
                        These packages will not be reported as unused
  --exclude EXCLUDE     A glob that will exclude paths otherwise matched
  --report {table,extended-table,line,none}
  --verbose, -v
  --quiet, -q

Configuration

All command line options are also supported as configuration options in a project's pyproject.toml. Options are specified under the [tool.bonded] table and have the same name and meaning as when specified as a command line option. The only option that will not have an effect when read from pyproject.toml is the pyproject setting.

For each setting, bonded will start with a default value, then override from options found in a pyproject.toml, if any, and finally override with options specified as arguments, if any.

An example entry:

[tool.bonded]
search_path = 'src/mypackage'
setup = 'src/setup.cfg'
exclude = ['__pycache__/']

Why can't it ..?

  • tell me the package I should depend on for undeclared modules?

    It is impossible for bonded to know what packages provide what modules if those packages are not installed locally. This is partly because python distribution names (what you download from pypi.org) and python package names (what you import) do not in any way have to relate to each other. It would be at best be wrong, at worst dangerous, to suggest you depend on packages based solely on name similarity.

  • figure out what modules a package supplies without it being installed locally?

    Bonded is not an environment manager, nor a package manager. Either of these tasks are independently complicated and best left to other tools that do them well. For the former try nox tox or hatch, for the latter try pip or hatch. Instead bonded is best used in conjunction with these tools.

  • use my virtualenv to figure out what my dependencies are?

    Declared dependencies are not equivalent to the contents of a virtualenv. Assuming that they are would remove bonded's ability to find the types of bugs where: someone installed it locally but didn't edit the metadata of the package, the dependency is only transitive and dependency requirements of other packages are being relied upon, a package is needlessly installed as there will be many packages required by the package's direct dependencies and installed locally but not required by the package itself.

  • read my setup.py?

    Anything can happen in a setup.py and bonded will not execute arbitrary code to find out a package's dependencies. Either move them to a declarative format, like setup.cfg, or tell bonded abut them explicitly with the --package option.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

bonded-0.5b1.tar.gz (1.8 MB view hashes)

Uploaded Source

Built Distribution

bonded-0.5b1-py2.py3-none-any.whl (17.3 kB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page