llm-code-lens 0.6.1

Ultra-fast intelligent code analysis tool for LLM context generation with Tree-sitter

🔍 CodeLens - Supercharge Your LLM Coding Experience

CodeLens is your AI coding assistant's best friend - an intelligent code analysis tool that transforms your codebase into LLM-optimized context. Stop wasting tokens on irrelevant files or struggling to explain your project structure. CodeLens does the heavy lifting, so you can focus on building great software.

---

🚀 Why CodeLens?

• Save time and tokens: Automatically extract the most relevant code context for your LLM
• Get better answers: Provide AI with structured insights about your codebase architecture
• Seamless LLM integration: One-click sharing with Claude, ChatGPT, Gemini and other LLMs
• Work smarter: Identify core files, entry points, and dependencies automatically
• Maintain with ease: Track TODOs, complexity hotspots, and technical debt

---

✨ Features

• Multi-language analysis: Deep insights for Python, JavaScript/TypeScript, and SQL codebases
• Direct LLM integration: Send analysis directly to Claude, ChatGPT, or Gemini with one click
• Smart code extraction: Identifies core files, entry points, and critical dependencies
• Interactive selection: Choose exactly which files and directories to analyze
• Token optimization: Splits large files into perfectly-sized chunks for LLM context windows
• Complexity metrics: Highlights complex functions and classes that need attention
• Maintenance tracking: Collects TODOs, FIXMEs, and technical debt indicators
• SQL database analysis: Examines stored procedures, views, and functions
• Pre-commit integration: Automatically runs tests before committing

---

📦 Installation

pip install llm-code-lens

That's it! No complex configuration needed.

---

🎮 Usage

Quick Start

Simply run:

llmcl

This launches the interactive interface where you can navigate your project, select files, and configure analysis options with just a few keystrokes.

Workflow

1. Select files: Navigate your project and choose what to analyze
2. Configure options: Set output format, LLM provider, and other settings
3. Run analysis: CodeLens examines your code and generates insights
4. Send to LLM: With one click, send everything to your preferred AI assistant

LLM Integration (Enhanced in v0.5.15!)

CodeLens now integrates directly with popular LLM providers:

llmcl --open-in-llm claude

Or select your provider in the interactive menu:

• Claude: Optimized format for Anthropic's Claude
• ChatGPT: Perfect context for OpenAI's models
• Gemini: Formatted for Google's Gemini
• None: Skip browser opening

When analysis completes, CodeLens:

1. Opens your chosen LLM in your default browser
2. Copies the complete analysis to your clipboard
3. Provides a system prompt optimized for code understanding

Just paste and start asking questions about your code!

New in v0.5.15: Enhanced AI Workflow

CodeLens now provides an optimized workflow for AI code assistance:

1. System Context: Automatically includes your OS, Python version, and architecture
2. Smart Editing: AI will ask for current files before suggesting changes (ensuring 100% accuracy)
3. Effortless Updates: All suggestions use search-and-replace format for quick implementation
4. Progress Tracking: Real-time progress bars show exactly what's happening during analysis

The enhanced system prompt ensures AI assistants provide more accurate, contextual suggestions that work perfectly with your specific development environment.

Interactive Interface

The intuitive terminal interface lets you:

• See real-time progress with detailed progress bars and file-by-file indicators
• Navigate with arrow keys (↑↓←→)
• Watch progress in real-time with animated progress bars and current file indicators
• Select files and directories with Space (includes all sub-elements)
• Switch sections with Tab
• Configure options with function keys (F1-F6)
• Confirm with Enter
• Cancel with Escape or Q

All your settings persist between runs, so you can quickly analyze the same files again.

Selection States

• [+] - Included file or directory
• [*] - Explicitly selected file or directory
• [-] - Excluded file or directory

Command Line Options

For CI/CD pipelines or scripting:

llmcl --format json --full --open-in-llm claude

Full options list:

• --output/-o: Output directory (default: .codelens)
• --format/-f: Output format (txt or json)
• --full: Export complete file contents
• --debug: Enable detailed logging
• --sql-server: SQL Server connection string
• --sql-database: Database to analyze
• --open-in-llm: LLM provider to open results in

---

📊 What You Get

CodeLens generates a comprehensive analysis in the .codelens directory:

1. Project Overview

• Total files, lines of code, and complexity metrics
• Visual project tree structure showing your codebase hierarchy
• Language distribution and project structure
• Entry points and core files identification
• System environment context (OS, Python version, architecture)

2. Smart Insights

• Architectural patterns detected
• Potential code smells and improvement areas
• Dependency relationships and import graphs

3. File-by-File Analysis

• Function and class inventories with complexity scores
• Documentation coverage and quality assessment
• TODOs and technical debt indicators

4. Full Content Export (Optional)

• Complete file contents split into token-optimized chunks
• Perfect for providing full context to your LLM

5. SQL Analysis (If Configured)

• Stored procedures, views, and functions inventory
• Object dependencies and relationships
• Parameter analysis and complexity metrics

---

🛠️ Configuration

Pre-commit Integration

Set up pre-commit hooks to ensure quality:

python scripts/install-hooks.py

This automatically runs tests before each commit.

SQL Server Configuration

Three ways to configure SQL analysis:

1. Environment Variables:

   export MSSQL_SERVER=your_server
   export MSSQL_DATABASE=your_database
   

2. Command Line:

   llmcl --sql-server "server_name" --sql-database "database_name"
   

3. Configuration File: Create sql-config.json and use:

   llmcl --sql-config sql-config.json
   

---

💡 Use Cases

For Developers

• Onboarding to new projects: Quickly understand unfamiliar codebases
• Refactoring planning: Identify complex areas that need attention
• Technical debt management: Track TODOs and maintenance needs
• Architecture discussions: Generate insights about code structure

For LLM Interactions

• Bug fixing: Provide perfect context for debugging issues
• Feature development: Help LLMs understand where and how to add features
• Code reviews: Get AI assistance with reviewing complex changes
• Documentation: Generate comprehensive docs from code analysis

For Teams

• Knowledge sharing: Create shareable insights about project structure
• Consistent context: Ensure everyone provides similar context to LLMs
• Codebase health: Track metrics over time to measure improvement
• SQL analysis: Understand database objects without direct access

---

🧩 SQL Server Integration

CodeLens provides deep analysis of SQL Server databases:

Prerequisites

• Microsoft ODBC Driver for SQL Server
• Appropriate database permissions

What You Get

• Complete inventory of stored procedures, views, and functions
• Parameter analysis and usage patterns
• Complexity metrics and dependency mapping
• Full object definitions with the --full flag

Security Best Practices

• Use environment variables for credentials
• Consider integrated security when possible
• Apply least-privilege principles for analysis

---

🔧 Development

Setting up the Environment

git clone https://github.com/SikamikanikoBG/codelens.git
cd codelens
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -e ".[dev]"

Running Tests

pytest

---

🤝 Contributing

We welcome contributions! To get started:

1. Fork the repository
2. Create a feature branch
3. Add tests for new functionality
4. Submit a pull request with a clear description

---

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.

---

🆘 Support

For issues or feature requests, please visit our GitHub Issues.

---

🌟 Star Us on GitHub!

If CodeLens has helped you, please consider giving us a star on GitHub. It helps others discover the tool and supports its continued development.