Skip to content

MatthewHambley/stylist

 
 

Repository files navigation

Stylist

BSD3 License GitHub release PyPI version Conda version GitHub merge testing

Stylist is a tool for checking code style. It implements a framework which supports multiple styles across multiple languages.

But aren't there many such tools out there, why create a new one?

The simple reason is that few of them support Fortran, a language still in widespread use in the scientific computing domain. They can also tend towards the zealous when enforcing the "one true style" whereas long running science models tend to diverge quite radically in what they like, style-wise.

Large models also tend to be implemented using more than one language so a tool which is capable of understanding all of them would be welcomed by science developers.

This is the goal of Stylist development.

The project is still in its infancy so only a few rules have been implemented and only for Fortran. Stubs are provided to show how it might be extended to support C.

Find the project at <https://github.com/MetOffice/stylist>

Installation

Installation can be as simple as pip install stylist or conda install -c conda-forge stylist.

As always it is also possible to install from the project source using pip install --editable .. The source may be obtained by downloading a tarball or by cloning the repository.

Usage

Stylist provides a command-line tool stylist for normal use. It can also be used as a package if you want to integrate it with another tool. Documentation regarding this second option is maintained in the project wiki.

On the Command Line

The command-line tool is not complicated to use:

stylist [-help] [-verbose]
        [-configuration FILENAME]
        [-style NAME]
        FILE ...

The only required arguments are a configuration file and one or more filenames. These are the files which will be checked. If a directory is specified then the tool will automatically descend into it checking all files which it recognises by extension.

If you want a running commentary of what the tool is doing then use the -verbose argument.

A configuration file may specified with -configuration . This file should be formatted as documented below. There must be a configuration file which defines at least one style. While the fallback mechanism which allows for user and site configurations has not been implemented it must be specified with this argument.

The configuration may define several styles, in which case one can be chosen using the -style argument. If it is not then the first in the configuration file will be used.

Configuration File

The configuration file is a simple python script where variable definitions are used to define the configuration.

Stylist uses the concept of "File Pipes" to specify the series of steps needed to process a particular file type. They are defined by creating a FilePipe object and associated with a file extension by assignment to a variable.

For example:

from stylist.source import (FilePipe,
                            FortranSource,
                            PFUnitProcessor,
                            FortranPreprocessor)

pf = FilePipe(FortranSource, PFUnitProcessor, FortranPreprocessor)

Here a pipeline consisting the pFUnit processor followed by the Fortran preprocessor produces Fortran source. This pipeline should be used for all files with the .pf extension.

Styles are collections of rules and are defined in a similar way. The variable name is the style name and the object is constructed with the rules which make up the style.

An illustrative example:

from re import compile as recompile
from stylist.style import Style
from stylist.fortran import FortranCharacterset, KindPattern

simple = Style(FortranCharacterset(),
               KindPattern(integer=recompile(r'i_.+'),
                           real=recompile(r'r_.+')))

About

Style checking for Fortran, PSyclone DSL, etc

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 94.1%
  • Fortran 5.9%