Skip to content

Core source code metrics development

Reading: 5 minutes, Last update: 10 March 2022

The following proposal has been initially prepared for the IUST Compiler and IUST Advanced Compiler courses in Winter/Spring 2022.

Note 1: Before reading this proposal ensure that you have read and understood the OpenUnderstand white-paper.

The goal of this proposal is to implement an extended version of the Sci-tools Understand APIs for computing source metrics. The following APIs are used to commute source code metrics at different abstraction levels: * understand.Db.metric * understand.Ent.metric

Students must form groups of up to four persons. Each group develops analysis passes to compute the value of source code metrics listed in Table 3.The exact list of the metrics will be assigned to each group subsequently. The description of the source code metrics listed in Table 3 may update during the semester.

The computation of source code metrics consists of two steps:

  • First, developing the required APIs for querying the database created in Phase 1 of the project. The APIs signature must follow the Sci-tools Understand Python API.
  • Second, querying the database using the developed API to find the appropriate entities and reference kinds involved in computing metrics according to the metric definitions.

Note 2: The computation of each source code metric must be implemented as a standalone Python module (single .py file) in the openuderstand.metrics package. Consider the existing modules in the packages as examples. Please strongly use the created database in Phase 1 to compute source code metrics.

Note 3: Each module should follow PEP -- Style Guide for Python Code and document well.

Note 4: Each group is asked to review the code of at least one other group in the classroom. Your final project score is computed considering the score given by referees to each member of the team.

Note 5: Before developing your codes ensure that you have pulled the latest version of the OpenUnderstand repository. For the final presentation, your code should work properly on the real-world software project existing in the benchmark directory of the OpenUnderstand repository.

Keep in touch!