StaticCodeMetrics

This repository provides helper-functions to compute static code metrics like, for instance, instability and abstractness of files and components. The development of these algorithms was inspired by the book Clean Architecture written by Robert C. Martin (https://www.amazon.com/dp/0134494164).

• Calculation of instability I: I = fan_out / (fan_in + fan_out), where fan_out denotes the number of outgoing dependencies from a file / component and fan_in the number of incoming dependencies to a file / component.

• Calculation of abstractness A: A = N_a / N_c, where N_a denotes the number of interfaces or abstract classes inside a file / component and N_c the number of classes inside a file / component.

• Distance D: D = |A + I - 1|, where a value of 0 indicates the file / component lies on the Main Sequence and a value of 1 denotes a file being far away of it.

• Main Sequence: A linear line going from point (0,1) to (1,0), where the coordinates (a,i) describe the value of Abstractness and the value of Instability, respectively. Both points given describe the most desireable points for a file / component: either maximal stable & abstract or maximal unstable & concrete. For all files / components given it is desireable to be on or close to the Main Sequence.

Installation

This application can either be installed via PyPi or by cloning this repository.

Requirements

Following packages are required for this application to be executed successfully:

• Python (>=3.7)
• numpy (>=1.18.3)
• pandas (>=1.0.3)
• matplotlib (>=3.2.1)

Simply via PyPi

Installing via PyPi is easy and straight forward. Executing following command will install this application for you:

$ pip install staticcodemetric-scm-pkg

Now the application can be started from the command line directly by running, for example,

$ staticcodemetric -h

Call Python-script directly

After cloning this repository, run _main_.py using Python from the root-directory, e.g.

$ python .\scm_modules\__main__.py -h

Usage

The static code checker can be started directly from the command line:

$ staticcodemetric -df <directory-path> -pl <programming-language> (-di | -ms) [-s] [-sp <save-path>]

Following options are available (required or optional):
-df <directory-path>: Path to the directory which contains the code-files to check. This directory will be processed recursively.
-pl <programming-language>: Programming language used in the files to check
-di: Plot distance metric
-ms: Plot Main Sequence
-s: Save computed metrics (either instability and abstractness or distance in default directory)
-sp <save-path>: Computed metrics are saved within provided path (but only if it exists)

Testing

Tests are written using Python's unittest library and can be locally executed using following commands from the root-directory:

$ python .\tests\get_test_modules.py
$ python .\tests\run_all_tests.py

Running the first script is required in order to copy all modules which are used for testing to a pre-defined folder. The tests which are executed by the latter script refer to this pre-defined folder. This workaround was necessary due to some dependency errors.

Development status

Currently, the metrics defined above can only be computed for C++ files.

Future development

In future, the metrics should be expanded to handle whole components (consisting of one or more files). Other programming languages will be tackled, too.