ejd 0.1.2

A CLI tool for static code analysis and fixing using Gemini AI

SDLC-Automation

Graduation project - CLI Backend for Static Issue Analysis and Fixing

Overview

This project provides a command-line interface (CLI) tool called ejd that helps developers analyze and fix static issues in their code using configurable rule sets and LLM integration.

Installation

pip install -e .

Usage

Basic Command Structure

ejd -si fix "path/to/your/file.py"

Interactive Workflow

1. Run the command with your source file path
2. The tool will prompt you to enter a rule set file path
3. Both files are validated and loaded into variables
4. The data is prepared for LLM processing

Example Usage

# Fix static issues in a Python file
ejd -si fix "src/main.py"

# The tool will then prompt:
# 📋 Please enter the path to the rule set file: rules/python_rules.txt

Features

• ✅ File validation for both source and rule set files
• 🔧 Interactive rule set path input with validation
• 💾 Automatic loading of file contents into variables
• 🤖 Ready-to-use data structure for LLM integration
• � Non-destructive workflow - creates fixed files while preserving originals
• �🚀 Extensible architecture for future enhancements

Gemini Flash 2.5 Integration

This tool is integrated with Google's Gemini Flash 2.5 model for intelligent static code analysis and automated fixing.

Setup

1. Install dependencies (automatically handled during installation):

   pip install -e .
   

2. Get your Gemini API key:

   • Visit: https://aistudio.google.com/app/apikey
   • Sign in with your Google account
   • Create a new API key
   • Copy the generated key

3. Configure the API key (choose one method):

   Option A: Using the setup script (Recommended)

   python setup_gemini.py
   

   Option B: Manual configuration

   # Create a .env file in the project root
   echo "GOOGLE_API_KEY=your_actual_api_key_here" > .env
   

   Option C: Environment variable

   # Windows PowerShell
   $env:GOOGLE_API_KEY="your_actual_api_key_here"
   
   # Windows CMD
   set GOOGLE_API_KEY=your_actual_api_key_here
   

How It Works

1. Analysis: Gemini analyzes your source code against the provided rules
2. Issue Detection: Identifies code quality, security, and performance issues
3. Automated Fixing: Generates corrected code that addresses all issues
4. Interactive Review: Shows you what issues were found and asks for confirmation before applying fixes

Example Analysis Results

The tool provides detailed analysis including:

• Issues Found: Line-by-line breakdown of problems
• Issue Types: Code quality, security vulnerabilities, performance issues
• Rule Violations: Which specific rules were violated
• Fixed Code: Complete corrected version of your code
• Best Practice Suggestions: Additional recommendations for improvement
• Markdown Report: Comprehensive analysis saved as a formatted markdown file

Markdown Report Features

Each analysis generates a detailed markdown report with:

• Structured Summary: Overview of all changes and improvements
• Issues Table: Organized table with line numbers, types, descriptions, and violated rules
• Complete Fixed Code: Full corrected source code in syntax-highlighted blocks
• Additional Suggestions: Best practice recommendations and improvement tips
• Timestamped Documentation: Generated date/time and source file information

Reports are automatically saved with timestamps (e.g., test_file_analysis_report_20251017_153045.md) for easy tracking and review.

Sample Output

🤖 Preparing to send data to Gemini Flash 2.5...
🔗 Testing connection to Gemini...
✅ Connected to Gemini successfully!
🎉 Analysis completed successfully!
📊 Issues found: 8
📄 Detailed markdown report saved to: test_file_analysis_report_20251017_153045.md

🐛 Issues identified:
   1. Line 2: Unused import 'sys' - Remove unused imports
   2. Line 15: SQL injection vulnerability - Use parameterized queries
   3. Line 25: Missing type hints - Add type annotations for better code clarity
   4. Line 30: Class name should be capitalized - Follow PEP 8 naming conventions
   5. Line 45: Use list comprehension for better performance

� Preview of fixed code (first 200 characters):
   # Fixed Python file - all static issues resolved
   import os
   import requests
   from typing import List, Optional
   import sqlite3
   
   def calculate_total(items: List[int]) -> int:...

�💡 Would you like to create a fixed version? (y/n): y
✅ Fixed version created: test_file-fixed.py
📁 Original file preserved: test_file.py
📝 Summary: Fixed 8 issues including security vulnerabilities and code quality improvements

💡 Additional suggestions:
   1. Consider adding docstrings to all functions and classes
   2. Use more descriptive variable names where possible
   3. Add error handling for file operations and user inputs

File Structure

ejad/
├── __init__.py
├── cli.py           # Main CLI entry point
└── static_issue.py  # Static issue fixing logic