Skip to content

tdresearchgroup/analyzeR-SonarQubePlugin

Repository files navigation

License: MIT GitHub Issues GitHub PRs DOI

analyzeR

analyzeR is an open-source SonarQube plugin for automated software analysis and measurement of R packages written using one (or a combination of) the following OO implementations in R: S3, S4, and/or R6. It is built on the SonarQube Plugin Example

The resulting measurements are presented through the SonarQube interface to improve readability and comparisons with other outputs.

Overall, analyzeR offers the following functionalities:

  • Enables automated static code analyses of OO-structured R packages, through a process that can be incorporated into a CI setup, following current industry practices1.

  • Calculates metrics on a per-file, per-module, per-class, and per-project basis. Note that, in OO R programming, projects/modules are represented as packages.

  • Capabilities to analyze R packages written with S3, S4 or R6 OO systems, with scope to extend the plugin to support additional OO systems, like proto and R.oo, or newer systems that may become available in the future

  • The plugin's implementation combines R, Python and Java. The original parsing of R packages is done through the RParser (a script written in R). The code scanning and measurement is written in Python and favour extensibility to add additional metrics. The RPlugin is written in Java. Documentation for all three parts is available here.

  • Measurement results are available through the SonarQube interface, respecting the placement and structure provided in the plugins used to examine other languages. This allows distinguishing between comparable and custom measurements, aggregation, and visualization of results.

Current Metrics

In version 1.0.0, support is only provided for S3, S4, and R6, by calculating traditional OO metrics2. At this point, analyzeR does not offer any R-specific metric, as the priority was to assemble a plugin compatible with current tools that could be easily extended.

The following metrics are considered:

Acron. Metric Description
LOC Lines of Code Number of non-blank lines
NPM # of Public Methods Number of public functions in an application
NOF # of Fields Number of public fields in an application
NMC # of Method Calls Number of function invocations
NMCI # of Internal Method Calls Number of function invocations of function defined in the application
NMCE # of External Method Calls Number of function invocations of function defined in other modules or packages
WMC Weighted Methods per Class Sum of the CycloComp of all functions in the application
AMC Average Method Complexity Average of the CycloComp of all functions in the application
RFC Response For a Class Number of functions that respond to the application itself
CBO Between Object Classes How many other modules an application (or module) is coupled to
CA Afferent Coupling How many other applications use the specific app./module
CE Efferent Coupling How many other modules are used by the specific app./module
MI Martin's Instability Indicates the necessity of performing modifications in an entity due to updates in other software entities.
LCOM Lack of Cohesion in Methods Difference between the number of function pairs without and with common non static fields
DAM Data Access Metrics Ratio of the number of private fields to total number of fields
NPRIF Number of Private Fields Number of private fields of an application or module
NPRIM Number of Private Methods Number of private functions of an application or module

Installation & Use

Please, refer to the current Wiki.

References

Footnotes

  1. 10.1109/SANER48275.2020.9054821

  2. 10.1145/3472163.3472172