repository·master·Indexed 24 days ago
https://github.com/rubik/radonA Python tool for computing source code metrics, including McCabe's cyclomatic complexity, Halstead metrics, the Maintainability Index, and raw metrics such as SLOC, comments, and blank lines. It provides a command-line interface, a programmatic API with three abstraction layers (Visitors, Helper Functions, and Harvesters), and a plugin for flake8.
Radon is a Python tool used to compute various code metrics. It supports the following types of analysis:
Cyclomatic Complexity (also called the McCabe number) represents the number of linearly independent paths through a block of code. It is calculated as the number of decisions in a code block plus 1. Radon computes this by analyzing the Python AST (Abstract Syntax Tree).
Common Python constructs and their effects on Cyclomatic Complexity (CC):
| Construct | Effect on CC | Reasoning |
|---|---|---|
if | +1 | Single decision |
elif | +1 | Adds another decision |
else | +0 | Decision is handled by the if |
case pattern | +1 | Single decision |
case _ | +0 | Decision is handled by the case pattern |
for | +1 | Decision at the start of the loop |
while | +1 | Decision at the while statement |
except | +1 | Each branch adds a new conditional path |
finally | +0 | Unconditionally executed |
with | +1 | Corresponds roughly to a try/except block |
assert | +1 | Internally behaves like a conditional |
| Comprehension | +1 | List/set/dict comprehensions or generator expressions act like for loops |
| Boolean Operator | +1 | Every and or or adds a decision point |
Radon's API is organized into three hierarchical layers depending on the level of abstraction required:
radon.visitors.ComplexityVisitor for cyclomatic complexity and radon.visitors.HalsteadVisitor for Halstead metrics.radon.complexity: For cyclomatic complexity analysis.radon.metrics: For Halstead metrics and the Maintainability Index (MI).radon.raw: For raw metrics like SLOC, LLOC, and LOC.radon.cli.Config object and a list of paths to analyze, and can export results to various formats (e.g., JSON or XML for cyclomatic complexity).Radon can be integrated into your workflow in two ways:
The Maintainability Index (MI) measures how easy it is to support and change source code. It is a factored formula involving SLOC (Source Lines of Code), Cyclomatic Complexity, and Halstead volume.
Radon uses a specific derivative of the MI formula, which is a hybrid of the SEI and Visual Studio implementations:
MI = max [ 0, 100 * (171 - 5.2 * ln(V) - 0.23 * G - 16.2 * ln(L) + 50 * sin(sqrt(2.4 * C))) / 171 ]
Where:
V: Halstead VolumeG: Total Cyclomatic ComplexityL: Number of Source Lines of Code (SLOC)C: Percent of comment lines (converted to radians)Note: Maintainability Index is considered an experimental metric and should be used with caution compared to other metrics.
You can install Radon using pip. If you are running Python versions earlier than 3.11 and wish to configure Radon using a pyproject.toml file, you must install the toml extra dependency.
To install the standard version:
$ pip install radonTo install with pyproject.toml support (for Python < 3.11):
$ pip install radon[toml]$ pip install radonTo use Radon's full CLI capabilities programmatically, use a Harvester (or specific harvesters like CCHarvester, RawHarvester, or MIHarvester).
radon.cli.Config object containing your desired configuration.Harvester with the config.Radon can inspect code metrics within .ipynb files. To use this feature, you must first install the nbformat package:
$ pip install nbformatWhen running Radon commands, use the following flags:
--include-ipynb: Enables scanning of Jupyter notebooks.--ipynb-cells: Enables reporting of metrics for individual cells.Note: Any % macros in the notebook will be ignored in the metrics.
Example command to run raw metrics on a notebook:
$ radon raw --include-ipynb --ipynb-cells .$ radon raw --include-ipynb --ipynb-cells .flake8 tool to report cyclomatic complexity during linting. To use this integration, you must have both radon and flake8 installed in your environment.RADONFILESENCODING environment variable to UTF-8 if your Python files contain Unicode characters to ensure they are analyzed correctly.Radon can analyze .ipynb files to inspect code metrics within Python cells. Any % macros are ignored during analysis.
Requirements:
Install the nbformat package: pip install nbformat.
Usage:
--include-ipynb flag to enable scanning of Jupyter notebooks.--ipynb-cells flag to report on individual cells within the notebooks.Radon uses specialized namedtuple subclasses to represent code blocks. These objects provide metadata about the location and complexity of functions and classes.
Represents a function or method block.
name: The identifier of the function.fullname: The qualified name (e.g., ClassName.method_name for methods, or just name for functions).letter: 'M' if it is a method, 'F' if it is a function.complexity: The cyclomatic complexity of the function.lineno, col_offset, endline: Location information.is_method (bool): Indicates if the function is a method.classname: The name of the class this method belongs to.closures: A list of functions defined within this function.Represents a class block.
name: The identifier of the class.fullname: The name of the class.letter: Always 'C'.complexity: The average complexity of the class (average complexity of its methods plus one).methods: A list of Function objects representing the class's methods.inner_classes: A list of Class objects defined within this class.lineno, col_offset, endline: Location information.