Code Smell Detector

A Python-based tool for detecting common code smells in Java source files. This tool can identify and report on six major code smell patterns that indicate potential issues with code maintainability and design.

Features

✨ Detects 6 Major Code Smells:

• Long Method - Methods that are excessively long
• God Class (Blob) - Classes with too many responsibilities
• Duplicated Code - Repeated code blocks that should be refactored
• Large Parameter List - Methods with too many parameters
• Magic Numbers - Hardcoded numeric values that should be constants
• Feature Envy - Methods overly interested in other classes' data

🔧 Flexible Configuration:

• YAML-based configuration with per-smell settings
• CLI overrides for runtime control
• Adjustable thresholds and detection parameters

📊 Multiple Output Formats:

• Detailed reports with line numbers and suggestions
• Summary reports with statistics
• JSON output for integration with other tools

Installation

Prerequisites

• Python 3.7+
• PyYAML package

Setup

# Install required dependencies
pip install PyYAML

# Clone or download the project
cd code-smell-detector/src

Usage

Basic Usage

# Analyze a single file
python detector_cli.py MyClass.java

# Analyze all Java files in a directory
python detector_cli.py src/

# List available detectors
python detector_cli.py --list-detectors

Advanced Usage

Selective Detection

# Only detect specific smells
python detector_cli.py src/ --only LongMethod,GodClass

# Exclude specific smells
python detector_cli.py src/ --exclude MagicNumbers,DuplicatedCode

Output Formats

# Detailed report (default)
python detector_cli.py src/ --format detailed

# Summary statistics
python detector_cli.py src/ --format summary

# JSON output
python detector_cli.py src/ --format json

Configuration

# Use custom configuration file
python detector_cli.py src/ --config custom_config.yaml

# Save report to file
python detector_cli.py src/ --output report.txt

# Verbose output
python detector_cli.py src/ --verbose

Configuration

The tool uses a YAML configuration file (config/config.yaml) to customize detection parameters:

code_smells:
  LongMethod:
    enabled: true
    threshold_lines: 30
  
  GodClass:
    enabled: true
    threshold_methods: 15
    threshold_lines: 200
  
  MagicNumbers:
    enabled: true
    exclude_common: true  # Skip 0, 1, -1, 2, etc.
    exclude_constants: true  # Skip final/static variables

CLI Precedence

• --only overrides everything else
• --exclude overrides config file
• Config file settings used as default

Code Smell Descriptions

1. Long Method

What it detects: Methods longer than the configured threshold (default: 30 lines) Why it matters: Long methods are harder to understand, test, and maintain Suggestion: Break into smaller, focused methods

2. God Class (Blob)

What it detects: Classes with too many methods or lines of code Why it matters: Violates Single Responsibility Principle Suggestion: Split into multiple smaller, focused classes

3. Duplicated Code

What it detects: Similar code blocks that appear multiple times Why it matters: Changes must be made in multiple places Suggestion: Extract into reusable methods

4. Large Parameter List

What it detects: Methods with many parameters (default: >5) Why it matters: Hard to use and understand Suggestion: Use parameter objects or builder pattern

5. Magic Numbers

What it detects: Hardcoded numeric values in code Why it matters: Unclear meaning and hard to maintain Suggestion: Extract into named constants

6. Feature Envy

What it detects: Methods that access other classes' data frequently Why it matters: Suggests misplaced responsibility Suggestion: Move method to appropriate class

Example Output

Detailed Report

============================================================
CODE SMELL DETECTION REPORT
============================================================
Total smells detected: 51
Active detectors: LongMethod, GodClass, DuplicatedCode, LargeParameterList, MagicNumbers, FeatureEnvy

🔍 LongMethod (1 occurrences)
----------------------------------------
📁 File: LibrarySystem.java
📍 Lines: 32-122
⚠️  Method 'borrowBook' is too long (90 lines, threshold: 30)
🔧 Severity: High
💡 Suggestion: Consider breaking this method into smaller, more focused methods

Summary Report

CODE SMELL SUMMARY
==============================
Total: 94 smells

By Type:
  FeatureEnvy: 7
  GodClass: 2
  LongMethod: 3
  MagicNumbers: 82

By Severity:
  High: 3
  Low: 82
  Medium: 9

Project Structure

code-smell-detector/
├── src/
│   ├── detector_cli.py           # Command-line interface
│   ├── detector_engine.py        # Main detection engine
│   └── detectors/
│       ├── base_detector.py      # Abstract base class
│       ├── structure_detectors.py # Long Method, God Class
│       ├── parameter_detectors.py # Large Parameter List, Magic Numbers
│       └── duplication_detectors.py # Duplicated Code, Feature Envy
├── config/
│   └── config.yaml              # Configuration file
├── test-files/                  # Sample Java files for testing
└── README.md                    # This documentation

Testing

The detector has been tested on a deliberately smelly Java library management system that contains all six code smell types. Test results show accurate detection and reporting.

Sample Test Commands

# Test all detectors on sample code
python detector_cli.py ../test-files/ --verbose

# Test specific detectors
python detector_cli.py ../test-files/LibrarySystem.java --only LongMethod,GodClass

# Generate JSON report
python detector_cli.py ../test-files/ --format json --output results.json

Extension Points

The tool is designed for extensibility:

1. Add New Detectors: Inherit from BaseDetector and implement the detection logic
2. Custom Output Formats: Extend the reporting system in detector_engine.py
3. Language Support: Adapt the parsing logic for other programming languages
4. IDE Integration: Use the JSON output format for integration with development tools

Contributing

To add a new code smell detector:

1. Create a new detector class inheriting from BaseDetector
2. Implement the detect() method and smell_type property
3. Add configuration options to config.yaml
4. Register the detector in detector_engine.py
5. Update documentation

License

This project is created for educational purposes as part of a Software Engineering assignment demonstrating code smell detection techniques.