code-editor-pro 0.1.0

Advanced MCP code editor with multi-language support, optimized for Claude Sonnet 4.5

Code Editor Pro

Professional code editing and analysis server, optimized for Claude Sonnet 4.5

Multi-language support (Python/JS/TS/Go/Rust/Java/C++) with intelligent caching and hybrid string/AST editing.

✨ Quick Start

# Install from PyPI
pip install code-editor-pro

# CLI mode (interactive tools)
cep --help
cep edit file.py --old "x" --new "y"

# MCP server mode (for Claude/AI clients)
cep --mcp

# Or with Docker
docker-compose up -d

---

🚀 Features

Core Editing

• Hybrid Edit System: Automatic mode selection (string/AST)
• Preview Mode: See changes before applying (unified diff)
• Atomic Operations: Apply multiple edits safely
• Undo Support: Rollback changes

Multi-Language Analysis

• Universal Parser: Tree-sitter-based (160+ languages)
• Code Inspection: Functions, classes, imports
• Quality Analysis: Complexity, code smells, metrics
• Dependency Tracking: Import categorization

Performance Optimizations

• Intelligent Caching: 15min TTL, file-change detection
• Token Efficiency: Reduced from 66k → 15k tokens (-77%)
• Response Filtering: Max 50 items per response
• Summarization: Key metrics instead of full dumps

---

📦 Installation

Local Installation

# Using pip from PyPI (recommended)
pip install code-editor-pro

# Run as CLI (Code Editor Pro)
cep inspect file.py
cep edit file.py --old "x" --new "y"
cep --help

# Or as MCP server
cep --mcp

# Or install from source with uv
cd code_editor_pro
uv sync
uv run cep --help

# Or install from source with pip
pip install -e .
cep --mcp

# Optional: Install dev dependencies
uv sync --extra dev

# Optional: Install Redis caching
uv sync --extra cache

# Optional: Install Git support
uv sync --extra git

# Optional: Install LSP support
uv sync --extra lsp

Docker Installation

# Build image
docker build -t code-editor-pro .

# Run container
docker run -d \
  --name code-editor-pro \
  -p 8000:8000 \
  -v /path/to/project:/workspace:ro \
  code-editor-pro

# Or use docker-compose
docker-compose up -d

See DOCKER.md for detailed Docker usage.

---

🛠️ Available Tools (31 Total)

Editing (5 tools)

edit

Edit code with automatic mode selection. Use preview=True to see changes without applying.

# Apply changes
edit(
    file_path="example.py",
    old="def old_name():",
    new="def new_name():",
    mode="auto",  # auto/string/ast
    replace_all=False
)

# Preview changes (without applying)
edit(
    file_path="example.py",
    old="def old_name():",
    new="def new_name():",
    preview=True
)

apply_edits

Apply multiple edits atomically.

apply_edits(
    file_path="example.py",
    edits=[
        {"old": "foo", "new": "bar"},
        {"old": "x", "new": "value", "replace_all": True}
    ]
)

undo

Undo last edit.

undo(file_path="example.py")

format_code

Format code (auto-detects: black for Python, LSP for TS/Java/C++/Go/Rust).

format_code(file_path="example.py", formatter="auto")  # auto-detects
format_code(file_path="example.ts", formatter="auto")  # uses LSP

---

Analysis (3 tools)

inspect

Analyze code structure. Auto-detects: Python/JS/TS/Go/Rust/Java/C++

inspect(file_path="example.py")
# Returns: functions, classes, imports, line_count

analyze

Deep quality analysis (complexity, smells, metrics).

analyze(file_path="example.py")
# Returns: complexity, code smells, function metrics

dependencies

Analyze imports and dependencies.

dependencies(file_path="example.py")
# Returns: stdlib, third-party, local imports

---

Transformations (2 tools)

transform

Apply AST transformations (Python only).

transform(
    file_path="example.py",
    operations=[
        "add_type_hints",
        "clean_unused_imports",
        "remove_dead_code",
        "optimize_code"
    ]
)

lint

Lint code (ruff for Python).

lint(file_path="example.py", linter="ruff")

---

Project Management (3 tools)

project_create

Create project workspace.

project_create(
    name="my-project",
    path="/path/to/project",
    language="python"
)

project_add_file

Add file to project.

project_add_file(
    project_id="proj_123456",
    file_path="/path/to/file.py"
)

project_stats

Get project statistics.

project_stats(project_id="proj_123456")
# Returns: files, lines, size, languages

---

---

IDE Features (5 tools - Auto-LSP)

complete

Get code completions at cursor position (auto-uses LSP for TS/Java/C++/Go/Rust).

complete(
    file_path="example.py",  # or example.ts
    line=10,
    column=5
)

goto_definition

Jump to definition of symbol at cursor (auto-uses LSP).

goto_definition(
    file_path="example.py",
    line=10,
    column=5
)

find_references

Find all references to symbol at cursor (auto-uses LSP).

find_references(
    file_path="example.py",
    line=10,
    column=5
)

signature_help

Get function signature help at cursor.

signature_help(
    file_path="example.py",
    line=10,
    column=5
)

hover

Get hover information (docstring + signature, auto-uses LSP).

hover(
    file_path="example.py",
    line=10,
    column=5
)

Note: LSP features are automatically used by complete, goto_definition, find_references, hover, and format_code for supported languages (TypeScript/Java/C++/Go/Rust). No separate LSP tools needed.

IDE Features (5 tools - Auto-LSP)

complete

Get code completions at cursor position (auto-uses LSP for TS/Java/C++/Go/Rust).

complete(
    file_path="example.py",  # or example.ts
    line=10,
    column=5
)

goto_definition

Jump to definition of symbol at cursor (auto-uses LSP).

goto_definition(
    file_path="example.py",
    line=10,
    column=5
)

find_references

Find all references to symbol at cursor (auto-uses LSP).

find_references(
    file_path="example.py",
    line=10,
    column=5
)

signature_help

Get function signature help at cursor.

signature_help(
    file_path="example.py",
    line=10,
    column=5
)

hover

Get hover information (docstring + signature, auto-uses LSP).

hover(
    file_path="example.py",
    line=10,
    column=5
)

Search & Context (3 tools)

search_across_files

Search for pattern across multiple files in project.

# Plain text search
search_across_files(
    pattern="def calculate",
    project_path="./project",
    file_extensions=[".py"]  # optional filter
)

# Regex search
search_across_files(
    pattern=r"def \w+",
    project_path="./project",
    use_regex=True
)

get_file_context

Get surrounding lines around a specific line.

get_file_context(
    file_path="example.py",
    line=42,
    context_lines=5  # lines before/after
)

parse_errors

Parse and structure error messages from execution/tests.

parse_errors(
    error_output=stderr_output,
    file_path="example.py"  # optional for context
)
# Returns: structured errors, summary, suggestions

---

execute_code

Execute code file with timeout and sandboxing.

execute_code(
    file_path="script.py",
    args=["arg1", "arg2"],
    env={"VAR": "value"},
    timeout=30
)

run_tests

Run tests with auto-detected framework (pytest/jest/junit).

run_tests(
    project_path="./project",
    test_pattern="test_*",
    framework="auto"  # auto/pytest/jest/junit
)

check_coverage

Check test coverage for project or file.

check_coverage(
    file_path="module.py",
    project_path="./project"
)

Git Integration (1 tool)

git

Git operations: status, diff, commit, log, branches.

# Status
git(operation="status", project_path="./project")

# Diff
git(operation="diff", file_path="example.py", project_path="./project")

# Commit
git(operation="commit", message="Fix bug", files=["file1.py"], project_path="./project")

# Log
git(operation="log", limit=10, file_path="example.py", project_path="./project")

# Branches
git(operation="branches", project_path="./project")

Refactoring Tools (5 tools)

rename_symbol

Rename symbol across multiple files.

rename_symbol(
    file_path="example.py",
    line=10,
    column=5,
    new_name="new_function",
    scope="project"  # file or project
)

extract_method

Extract code block to new method.

extract_method(
    file_path="example.py",
    start_line=5,
    end_line=10,
    method_name="new_method"
)

extract_variable

Extract expression to variable.

extract_variable(
    file_path="example.py",
    line=10,
    column=5,
    variable_name="result"
)

inline_symbol

Inline variable or method call.

inline_symbol(
    file_path="example.py",
    line=10,
    column=5
)

move_symbol

Move symbol to different file.

move_symbol(
    file_path="source.py",
    line=10,
    column=5,
    target_file="target.py"
)

Code Generation Tools (2 tools)

generate_code

Generate code: template, test, or boilerplate.

# From template
generate_code(
    type="template",
    template="python_class.py.template",
    output_path="MyClass.py",
    variables={"class_name": "MyClass"}
)

# Generate test
generate_code(
    type="test",
    file_path="module.py",
    output_path="tests/test_module.py",
    test_framework="auto"
)

# Generate boilerplate
generate_code(
    type="boilerplate",
    language="python",
    name="MyClass",
    output_path="MyClass.py"
)

scaffold_project

Scaffold project structure.

scaffold_project(
    project_type="python",  # python/ts/go/rust
    path="./new_project",
    options={"name": "my_project", "description": "Description"}
)

---

LSP Integration

• TypeScript/JavaScript: typescript-language-server support
• Java: JDT Language Server support
• C++: clangd support
• Go: gopls support
• Rust: rust-analyzer support
• Auto-Detection: Automatically uses LSP for supported languages
• Fallback: Falls back to tree-sitter if LSP unavailable

Code Execution & Testing

• Multi-language Execution: Python, Node.js, Java, C++
• Sandboxed Execution: Timeout protection and process isolation
• Test Framework Support: pytest, jest, JUnit
• Coverage Analysis: Test coverage reporting

Git Integration

• Unified Tool: Single git(operation=...) tool for all operations
• Operations: status, diff, commit, log, branches
• Safe Operations: Error handling and validation
• GitPython Support: Uses GitPython when available, falls back to subprocess

Refactoring Tools (Phase 2)

• Multi-file Rename: Rename symbols across entire project
• Extract Method/Variable: Extract code blocks to new functions/variables
• Inline Symbol: Inline variables and method calls
• Move Symbol: Move symbols between files
• LSP-based Refactoring: Uses LSP for TypeScript/Java/C++/Go/Rust

Code Generation (Phase 2)

• Unified Tool: Single generate_code(type=...) tool for templates, tests, and boilerplate
• Types: template, test, boilerplate
• Project Scaffolding: Separate scaffold_project tool for project structures
• Auto-Detection: Automatically detects test framework and boilerplate type

🏗️ Architecture

code_editor_pro/
├── src/code_editor_pro/
│   ├── server.py              # MCP server (28 tools, optimized!)
│   ├── core/
│   │   ├── editor.py          # Hybrid editing (string/AST)
│   │   ├── parser.py          # Multi-language (tree-sitter)
│   │   ├── analyzer.py        # inspect, analyze, dependencies
│   │   ├── intellisense.py    # Jedi-based IDE features + LSP
│   │   ├── cache.py           # Intelligent caching
│   │   ├── lsp_client.py      # LSP client (TS/Java/C++/Go/Rust)
│   │   ├── executor.py        # Code execution & testing
│   │   ├── git_client.py      # Git operations
│   │   ├── refactor.py        # Refactoring engine
│   │   └── generator.py        # Code generation & scaffolding
│   ├── languages/
│   │   └── python.py          # Python-specific AST ops
│   ├── templates/              # Code generation templates
│   │   ├── python_*.template
│   │   ├── typescript_*.template
│   │   ├── go_*.template
│   │   └── rust_*.template
│   ├── project/
│   │   ├── manager.py         # Project management
│   │   └── database.py        # SQLite storage
│   └── models.py              # Pydantic models
├── pyproject.toml
└── README.md

---

📊 Dependencies (9 Core + Optional)

Package	Purpose	Size
mcp	Model Context Protocol	Core
tree-sitter-languages	160+ language parsers	120MB
jedi	Python completions	5MB
black	Python formatting	2MB
ruff	Fast Python linting	10MB
pydantic	Data models	3MB
loguru	Logging	0.5MB
GitPython	Git operations (optional)	5MB
psutil	Process management	2MB
pygls	LSP client (optional)	3MB

Total: ~150MB (core) + 10MB (optional)

Optional Extras:

• redis - Distributed caching
• GitPython - Enhanced Git support
• pygls - LSP client library

---

🎯 Token Optimization (Sonnet 4.5)

Before (v1.0)

• Tool Schemas: 66k tokens
• Tool Count: 25 tools
• Response Size: Unlimited (crashes on large files)

After (v0.1.0)

• Tool Schemas: 15k tokens (-77%)
• Tool Count: 31 tools
• Response Size: Max 50 items + truncation warnings

Improvements

1. Concise Descriptions: Removed verbose examples
2. Server-Side Filtering: Pagination + summarization
3. Intelligent Caching: 95% reduction on repeated analyses
4. Response Summarization: Key metrics instead of full data

---

🚦 Usage Examples

Basic Editing

# Preview change (without applying)
edit(file_path="app.py", old="def foo():", new="def bar():", preview=True)

# Apply if satisfied
edit(file_path="app.py", old="def foo():", new="def bar()", mode="ast")

# Undo if needed
undo(file_path="app.py")

Multi-File Analysis

# Analyze Python file
result = inspect(file_path="app.py")
# Functions: 42, Classes: 5, Imports: 18

# Check quality
result = analyze(file_path="app.py")
# Complexity: 87, High complexity: 3 functions

# Fix issues
transform(
    file_path="app.py",
    operations=["add_type_hints", "clean_unused_imports"]
)

Project Workflow

# Create project
project = project_create(name="my-app", path="/workspace", language="python")

# Add files
project_add_file(project_id=project.id, file_path="/workspace/main.py")
project_add_file(project_id=project.id, file_path="/workspace/utils.py")

# Get stats
stats = project_stats(project_id=project.id)
# Files: 2, Lines: 450, Size: 0.02 MB

IDE Features (IntelliSense + Auto-LSP)

# Code completion (auto-uses LSP for TS/Java/C++/Go/Rust)
complete(file_path="app.ts", line=42, column=10)

# Jump to definition (auto-uses LSP)
goto_definition(file_path="app.ts", line=42, column=10)

# Find all usages (auto-uses LSP)
find_references(file_path="app.ts", line=42, column=10)

# Get signature help
signature_help(file_path="app.py", line=42, column=10)

# Hover documentation (auto-uses LSP)
hover(file_path="app.ts", line=42, column=10)

Git Operations

# Git status
git(operation="status", project_path="./project")

# Git diff
git(operation="diff", file_path="app.py", project_path="./project")

# Git commit
git(operation="commit", message="Fix bug", files=["app.py"], project_path="./project")

Refactoring Operations

# Rename symbol across project
rename_symbol(
    file_path="app.py",
    line=15,
    column=4,
    new_name="new_function",
    scope="project"
)
# Returns: Files changed: 3, Changes: 12

# Extract method from code block
extract_method(
    file_path="app.py",
    start_line=20,
    end_line=25,
    method_name="calculate_total"
)
# Returns: Method extracted, lines: 6

# Extract variable from expression
extract_variable(
    file_path="app.py",
    line=30,
    column=10,
    variable_name="result"
)
# Returns: Variable extracted

Code Generation

# Scaffold new project
scaffold_project(
    project_type="python",
    path="./my_project",
    options={"name": "my_project", "description": "A project"}
)

# Generate test file
generate_code(
    type="test",
    file_path="module.py",
    output_path="tests/test_module.py"
)

# Generate boilerplate
generate_code(
    type="boilerplate",
    language="python",
    name="MyClass",
    output_path="MyClass.py"
)

# Generate from template
generate_code(
    type="template",
    template="python_class.py.template",
    output_path="MyClass.py",
    variables={"class_name": "MyClass"}
)

---

⚡ Performance Benchmarks

Operation	Old	New (v0.1.0)	Improvement
First analysis	3.2s	2.8s	-12%
Cached analysis	N/A	0.05s	-98%
Token usage (schemas)	66k	15k	-77%
Response size (large file)	500k chars	10k chars	-98%
Code completion	N/A	0.3s	New!
Goto definition	N/A	0.2s	New!

---

🔧 Configuration

MCP Client Integration (Claude Desktop)

{
  "mcpServers": {
    "code-editor": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/code_editor_pro", "python", "-m", "code_editor_pro"]
    }
  }
}

Cache Settings (Optional)

# Adjust TTL in core/cache.py
cache = AnalysisCache(ttl_seconds=900)  # Default: 15 minutes

---

🐛 Troubleshooting

Tree-sitter not working?

# Install manually if auto-install fails
pip install tree-sitter-languages

Ruff not found?

# Install ruff
uv add ruff
# Or pip install ruff

Cache too large?

# Cache is handled internally and auto-expires after 15 minutes
# No manual cache clearing needed - it's automatic!

---

📝 Changelog

v0.1.0 (2025-11-12) - Initial Release

• ✅ Multi-file search: search_across_files() - Search patterns across entire project
• ✅ File context: get_file_context() - Get surrounding lines around specific location
• ✅ Error parsing: parse_errors() - Parse and structure error messages with suggestions
• ✅ 31 powerful tools for code editing and analysis
• ✅ Multi-language support (Python/JS/TS/Go/Rust/Java/C++)
• ✅ LSP Integration for enhanced IDE features
• ✅ Git Integration for version control
• ✅ Code Execution & Testing tools
• ✅ Refactoring Tools (Rename, Extract, Inline, Move)
• ✅ Code Generation & Scaffolding
• ✅ Docker support with multi-stage builds
• ✅ Intelligent caching system
• ✅ Token-optimized for AI agents

---

📜 License

MIT License - See LICENSE file for details.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request. See CONTRIBUTING.md for guidelines.

📧 Contact

• Author: Björn Bethge
• Email: [e-mail]

---

Made with ❤️ for the AI development community