codemap 0.2.0rc1

CLI tool that generates optimized markdown docs and streamline Git workflows

CodeMap

[!Caution] CodeMap is currently in active development. Use with caution and expect breaking changes.

Overview

CodeMap is an AI-powered developer toolkit. Generate optimized docs, analyze code semantically, and streamline Git workflows (AI commits, PRs) with multi-LLM support via an interactive CLI.

Features

• 🎯 Token-optimized documentation generation
• 📝 Rich markdown output with code structure
• 🌳 Repository structure visualization
• 🔄 Smart Git commit assistance with AI-generated messages
• 🔃 AI-powered PR creation and management
• 💬 Ask questions about your codebase using RAG
• 🔍 Index and search your repository for semantic understanding

Installation

[!Important] CodeMap currently only supports Unix-based platforms (macOS, Linux). For Windows users, we recommend using Windows Subsystem for Linux (WSL).

[!Tip] After installation, you can use either codemap or the shorter alias cm to run the commands.

Installation using pipx (Recommended)

Using pipx is recommended as it installs the package in an isolated environment and automatically manages the PATH.

# Ensure pipx is installed (install it if you haven't)
# python3 -m pip install --user pipx
# python3 -m pipx ensurepath

# Install codemap from PyPI
pipx install codemap

Alternative: Manual Installation using pip

If you prefer not to use pipx, you can install using pip directly:

# Install with pip (user installation)
pip install --user codemap

# Make sure your PATH includes the user bin directory
# Add the following to your shell profile (e.g., ~/.bashrc, ~/.zshrc):
# export PATH="$HOME/.local/bin:$PATH"
# Or find the correct path using: python3 -m site --user-base

Development Version (Latest GitHub)

If you want to try the latest development version with unreleased features:

# Using pipx
pipx install git+https://github.com/SarthakMishra/codemap.git

# Or using pip
pip install --user git+https://github.com/SarthakMishra/codemap.git

Updating CodeMap

To update CodeMap to the latest version:

# If installed with pipx from PyPI
pipx upgrade codemap

# If installed with pip from PyPI
pip install --user --upgrade codemap

# If installed from GitHub
pipx upgrade codemap  # or
pip install --user --upgrade git+https://github.com/SarthakMishra/codemap.git

Uninstalling

# If installed with pipx
pipx uninstall codemap

# If installed with pip
pip uninstall codemap

Index Repository Command

Index and watch repositories for efficient vector-based search and semantic understanding. This command processes your codebase into a vector database for use with other features like the ask command.

Command Options

codemap index [PATH] [OPTIONS]

Arguments:

• PATH: Path to the repository root directory (defaults to current directory)

Options:

• --sync/--no-sync: Synchronize the vector database with the current Git state on startup (default: sync)
• --watch, -w: Keep running and watch for file changes, automatically syncing the index
• --log-level, -L: Set the logging level (e.g., DEBUG, INFO, WARNING)
• --verbose, -v: Enable verbose logging
• --config, -c: Path to custom configuration file

Examples

# Index the current directory (one-time)
codemap index
# Or using the alias:
cm index

# Index with file watching enabled
codemap index --watch

# Skip initial sync but enable watching
codemap index --no-sync --watch

# Index a specific repository with verbose output
codemap index /path/to/repo --verbose

# Set specific logging level
codemap index --log-level DEBUG

Ask Questions About Your Codebase

Use the ask command to query your codebase using natural language and get AI-powered answers based on your code's content and structure.

Command Options

codemap ask [QUESTION] [OPTIONS]

Arguments:

• QUESTION: Your question about the codebase (omit for interactive mode)

Options:

• --path, -p: Path to the repository root (defaults to current directory)
• --model: LLM model to use (e.g., 'openai/gpt-4o-mini'). Overrides config.
• --api-base: Override the LLM API base URL
• --api-key: Override the LLM API key (use environment variables for security)
• --interactive, -i: Start an interactive chat session
• --verbose, -v: Enable verbose logging

Examples

# Ask a single question
codemap ask "How does the configuration loader work?"
# Or using the alias:
cm ask "What does the file watcher module do?"

# Start an interactive chat session
codemap ask --interactive

# Specify a repository path
codemap ask "Explain the vector database schema" -p /path/to/repo

# Use a specific LLM model
codemap ask "What is the purpose of the ProcessingPipeline?" --model anthropic/claude-3-sonnet

# Enable verbose logging
codemap ask "List all available CLI commands" -v

Generate Markdown Docs

Generate optimized markdown documentation and directory structures for your project:

Command Options

codemap gen [PATH] [OPTIONS]

Arguments:

• PATH: Path to the codebase to analyze (defaults to current directory)

Options:

• --output, -o: Output file path for the documentation (overrides config)
• --config, -c: Path to custom configuration file
• --max-content-length: Maximum content length for file display (set to 0 for unlimited, overrides config)
• --lod: Level of Detail for code analysis (signatures, structure, docs, full). Default: docs. Overrides config.
• --semantic/--no-semantic: Enable/disable semantic analysis using LSP. Default: enabled. Overrides config.
• --tree/--no-tree: Include/exclude directory tree in output. Overrides config (gen.include_tree).
• --verbose, -v: Enable verbose logging
• --process/--no-process: Process the codebase before generation. Default: enabled.
• --entity-graph/--no-entity-graph: Include/exclude entity relationship graph (Mermaid) in output. Overrides config (gen.include_entity_graph).
• --mermaid-entities: Comma-separated list of entity types (e.g., 'module,class,function'). Overrides config (gen.mermaid_entities).
• --mermaid-relationships: Comma-separated list of relationship types (e.g., 'declares,imports,calls'). Overrides config (gen.mermaid_relationships).
• --mermaid-legend/--no-mermaid-legend: Show/hide the legend in the Mermaid diagram. Overrides config (gen.mermaid_show_legend).
• --mermaid-unconnected/--no-mermaid-unconnected: Remove/keep nodes with no connections in the Mermaid diagram. Overrides config (gen.mermaid_remove_unconnected).

Examples

# Generate documentation for current directory using defaults
codemap gen
# Or using the alias:
cm gen

# Generate for a specific path with full detail and no semantic analysis
codemap gen /path/to/project --lod full --no-semantic

# Generate docs with signatures only and custom Mermaid settings
cm gen --lod signatures --mermaid-entities "class,function" --mermaid-relationships "calls"

# Generate only directory tree (implicitly disables entity graph)
codemap gen --tree --no-entity-graph

# Custom output location and content length
codemap gen -o ./docs/codebase.md --max-content-length 1500

# Use custom configuration file
codemap gen --config custom-config.yml

# Verbose mode for debugging
codemap gen -v

Smart Commit Feature

Create intelligent Git commits with AI-assisted message generation. The tool analyzes your changes, splits them into logical chunks, and generates meaningful commit messages using LLMs.

Basic Usage

# Basic usage with default settings (interactive, semantic splitting)
codemap commit
# Or using the alias:
cm commit

# Commit with a specific message (skips AI generation)
codemap commit -m "feat: add new feature"

# Commit all changes (including untracked files)
codemap commit -a

# Use a specific LLM model
codemap commit --model groq/llama-3.1-8b-instant

# Bypass git hooks (e.g., pre-commit)
codemap commit --bypass-hooks

Command Options

codemap commit [PATH] [OPTIONS]

Arguments:

• PATH: Path to repository or specific file to commit (defaults to current directory)

Options:

• --message, -m: Specify a commit message directly (skips AI generation)
• --all, -a: Commit all changes (stages untracked files)
• --model: LLM model to use for message generation (default: openai/gpt-4o-mini). Overrides config (commit.llm.model).
• --strategy, -s: Strategy for splitting diffs (default: semantic). Options: file, hunk, semantic. Overrides config (commit.strategy).
• --non-interactive: Run in non-interactive mode (accepts all generated messages)
• --bypass-hooks: Bypass git hooks with --no-verify (overrides config commit.bypass_hooks).
• --verbose, -v: Enable verbose logging

Interactive Workflow

The commit command provides an interactive workflow that:

1. Analyzes your changes and splits them into logical chunks
2. Generates AI-powered commit messages for each chunk
3. Allows you to:
   • Accept the generated message
   • Edit the message before committing
   • Regenerate the message
   • Skip the chunk
   • Exit the process

Commit Linting Feature

CodeMap includes automatic commit message linting to ensure your commit messages follow conventions:

1. Automatic Validation: Generated commit messages are automatically validated against conventional commit standards.

2. Linting Rules:

   • Type must be one of the allowed types (configurable in .codemap.yml)
   • Type must be lowercase
   • Subject must not end with a period
   • Subject must be at least 10 characters long
   • Header line should not exceed the configured maximum length (default: 72 characters)
   • Scope must be in lowercase (if provided)
   • Header must have a space after the colon
   • Description must start with an imperative verb

3. Auto-remediation: If a generated message fails linting, CodeMap will:

   • Identify the specific issues with the message
   • Automatically attempt to regenerate a compliant message (up to 3 attempts)
   • Provide feedback during regeneration with a loading spinner

4. Fallback Mechanism: If all regeneration attempts fail, the last message will be used, with linting status indicated.

Commit Strategy

The tool uses semantic analysis to group related changes together based on:

• File relationships (e.g., implementation files with their tests)
• Code content similarity
• Directory structure
• Common file patterns

[!Note] The semantic strategy utilizes a custom, distilled version of the Qodo/Qodo-Embed-1-1.5B model, named Qodo-Embed-M-1-1.5B-M2V-Distilled. This Model2Vec distilled model is significantly smaller (233MB vs 5.9GB) and faster (~112x) than the original while retaining ~85% of its performance, making semantic analysis efficient. You can find more details here.

Environment Variables

The following environment variables can be used to configure the commit command:

• OPENAI_API_KEY: OpenAI API key (default LLM provider)
• ANTHROPIC_API_KEY: Anthropic API key
• GROQ_API_KEY: Groq API key
• MISTRAL_API_KEY: Mistral API key
• COHERE_API_KEY: Cohere API key
• TOGETHER_API_KEY: Together API key
• OPENROUTER_API_KEY: OpenRouter API key

Examples

# Basic interactive commit
codemap commit

# Commit specific files
codemap commit path/to/file.py

# Use a specific model with custom strategy
codemap commit --model anthropic/claude-3-sonnet --strategy semantic

# Non-interactive commit with all changes
codemap commit -a --non-interactive

# Commit with verbose logging
codemap commit -v

# Demonstrate automatic linting and regeneration
codemap commit --verbose  # Will show linting feedback and regeneration attempts

PR Command Feature

The codemap pr command helps you create and manage pull requests with ease. It integrates with the existing codemap commit command to provide a seamless workflow from code changes to pull request creation.

PR Command Features

• Create branches with intelligent naming based on your current changes
• Support for multiple Git workflow strategies (GitHub Flow, GitFlow, Trunk-Based)
• Rich branch visualization with metadata and relationships
• Smart base branch selection based on branch type
• Automatic content generation for different PR types (feature, release, hotfix)
• Workflow-specific PR templates based on branch type
• Interactive PR content editing with previews
• Update existing PRs with new commits
• Configurable via .codemap.yml for team-wide settings

PR Command Requirements

• Git repository with a remote named origin
• GitHub CLI (gh) installed for PR creation and management
• Valid GitHub authentication for the gh CLI

Creating a PR

codemap pr create [PATH] [OPTIONS]
# Or using the alias:
cm pr create [PATH] [OPTIONS]

Arguments:

• PATH: Path to the codebase to analyze (defaults to current directory)

Options:

• --branch, -b: Target branch name
• --type, -t: Branch type (e.g., feature, release, hotfix, bugfix). Valid types depend on workflow strategy.
• --base: Base branch for the PR (defaults to repo default or workflow-defined default)
• --title: Pull request title
• --desc, -d: Pull request description (file path or text)
• --no-commit: Skip the commit process before creating PR
• --force-push, -f: Force push the branch
• --workflow, -w: Git workflow strategy (github-flow, gitflow, trunk-based). Overrides config (pr.strategy).
• --non-interactive: Run in non-interactive mode
• --model, -m: LLM model for content generation (overrides config llm.model).
• --verbose, -v: Enable verbose logging

Updating a PR

codemap pr update [PATH] [OPTIONS]
# Or using the alias:
cm pr update [PATH] [OPTIONS]

Arguments:

• PATH: Path to the codebase to analyze (defaults to current directory)

Options:

• --pr: PR number to update (required if not updating PR for current branch)
• --title: New PR title
• --desc, -d: New PR description (file path or text)

[!Note] --no-commit is NOT an option for 'update'

• --force-push, -f: Force push the branch (use with caution)
• --non-interactive: Run in non-interactive mode
• --verbose, -v: Enable verbose logging

Git Workflow Strategies

The PR command supports multiple Git workflow strategies:

1. GitHub Flow (default)

   • Simple, linear workflow
   • Feature branches merge directly to main

2. GitFlow

   • Feature branches → develop
   • Release branches → main
   • Hotfix branches → main (with back-merge to develop)

3. Trunk-Based Development

   • Short-lived feature branches
   • Emphasizes small, frequent PRs

PR Template System

CodeMap includes a robust PR template system that automatically generates appropriate titles and descriptions based on:

1. The selected workflow strategy (GitHub Flow, GitFlow, Trunk-Based)
2. The branch type (feature, release, hotfix, bugfix)
3. The changes being made

Workflow-Specific Templates

Each Git workflow strategy provides specialized templates:

GitHub Flow Templates

• Simple, general-purpose templates
• Focus on changes and testing
• Example format: {description} for title, structured sections for description

GitFlow Templates

• Specialized templates for each branch type:
  • Feature: Focus on new functionality with implementation details
  • Release: Structured release notes with features, bug fixes, and breaking changes
  • Hotfix: Emergency fix templates with impact analysis
  • Bugfix: Templates focused on bug description, root cause, and testing

Trunk-Based Templates

• Concise templates for short-lived branches
• Focus on quick implementation and rollout plans
• Emphasis on testing and deployment strategies

Template Configuration

In your .codemap.yml, you can configure how templates are used:

pr:
  # Content generation settings
  generate:
    title_strategy: "template"  # Options: commits, llm, template
    description_strategy: "template"  # Options: commits, llm, template
    use_workflow_templates: true  # Use workflow-specific templates (default: true)
    
    # Custom template (used when use_workflow_templates is false)
    description_template: |
      ## Changes
      {description}
      
      ## Testing
      - [ ] Unit tests
      - [ ] Integration tests
      
      ## Additional Notes
      
      ## Related Issues
      Closes #

Configuration Options:

• title_strategy: How PR titles are generated

  • commits: Generate from commit messages
  • llm: Use AI to generate titles
  • template: Use workflow-specific templates

• description_strategy: How PR descriptions are generated

  • commits: Generate structured content from commit messages
  • llm: Use AI to generate descriptions
  • template: Use workflow-specific templates

• use_workflow_templates: Whether to use built-in templates for each workflow strategy

  • When true: Uses the appropriate template based on workflow and branch type
  • When false: Uses the custom template defined in description_template

• description_template: Custom template with placeholder variables

  • {description}: Brief description of changes
  • {changes}: List of changes from commits
  • {user}: Current Git user
  • Supports any Markdown formatting

Examples

# Create PR using workflow-specific templates (GitFlow)
codemap pr create --workflow gitflow --type feature

# Create PR with custom title but workflow-based description
codemap pr create --title "My Custom Title" --workflow trunk-based

# Override both the workflow template and use custom description
codemap pr create --desc "Custom description with **markdown** support"

# Non-interactive PR creation with defined template usage
codemap pr create --non-interactive --workflow gitflow --type release

LLM Provider Support

CodeMap supports multiple LLM providers through LiteLLM:

# Using OpenAI (default)
codemap commit --model openai/gpt-4o-mini
# Or using the alias:
cm commit --model openai/gpt-4o-mini

# Using Anthropic
codemap commit --model anthropic/claude-3-sonnet-20240229

# Using Groq (recommended for speed)
codemap commit --model groq/llama-3.1-8b-instant

# Using OpenRouter
codemap commit --model openrouter/meta-llama/llama-3-8b-instruct

Configuration

Create a .codemap.yml file in your project root to customize the behavior. Below are all available configuration options with their default values from config.py:

# LLM configuration (applies globally unless overridden by command-specific LLM config)
llm:
  model: openai/gpt-4o-mini  # Default LLM model (provider/model_name format)
  api_base: null             # Custom API base URL (e.g., for local LLMs or proxies)

# Documentation Generation Settings ('gen' command)
gen:
  max_content_length: 5000       # Max content length per file (0 = unlimited)
  use_gitignore: true            # Respect .gitignore patterns
  output_dir: documentation       # Directory for generated docs
  include_tree: true             # Include directory tree in output
  include_entity_graph: true     # Include Mermaid entity relationship graph
  semantic_analysis: true        # Enable semantic analysis using LSP
  lod_level: docs                # Level of Detail: signatures, structure, docs, full
  mermaid_entities:              # Entity types for Mermaid graph
    - module
    - class
    - function
    - method
    - constant
    - variable
    - import
  mermaid_relationships:         # Relationship types for Mermaid graph
    - declares
    - imports
    - calls
  mermaid_show_legend: true      # Show legend in Mermaid diagram
  mermaid_remove_unconnected: false # Remove unconnected nodes in Mermaid diagram

# Processor configuration (background analysis)
processor:
  enabled: true                  # Enable background processing (currently unused)
  max_workers: 4                 # Max parallel workers for analysis
  ignored_patterns:              # Patterns to ignore during analysis
    - "**/.git/**"
    - "**/__pycache__/**"
    - "**/.venv/**"
    - "**/node_modules/**"
    - "**/*.pyc"
    - "**/dist/**"
    - "**/build/**"
  default_lod_level: signatures  # Default LOD for background processing

# Ask Command Configuration ('ask' command)
ask:
  interactive_chat: false        # Start an interactive chat session by default
  model: null                    # LLM model to use (falls back to global llm.model)
  
# Index Command Configuration ('index' command)
watcher:
  debounce_delay: 2.0            # Delay in seconds for file watcher to debounce changes

# Commit Feature Configuration ('commit' command)
commit:
  strategy: semantic             # Diff splitting strategy: file, hunk, semantic
  bypass_hooks: false            # Default for --bypass-hooks flag (--no-verify)
  # Note: 'commit.llm' config is deprecated; use global 'llm' section

  convention:                    # Commit convention settings (based on Conventional Commits)
    types:                       # Allowed commit types
      - feat
      - fix
      - docs
      - style
      - refactor
      - perf
      - test
      - build
      - ci
      - chore
    scopes: []                   # Optional scopes (can be auto-derived if empty)
    max_length: 72               # Max length for commit subject line

  lint:                          # Commitlint rule configuration (see https://commitlint.js.org/#/reference-rules)
    # Header rules
    header_max_length: { level: ERROR, rule: always, value: 100 }
    header_case: { level: DISABLED, rule: always, value: lower-case }
    header_full_stop: { level: ERROR, rule: never, value: . }
    # Type rules
    type_enum: { level: ERROR, rule: always } # Uses types from commit.convention.types
    type_case: { level: ERROR, rule: always, value: lower-case }
    type_empty: { level: ERROR, rule: never }
    # Scope rules
    scope_case: { level: ERROR, rule: always, value: lower-case }
    scope_empty: { level: DISABLED, rule: never }
    scope_enum: { level: DISABLED, rule: always } # Uses scopes from commit.convention.scopes
    # Subject rules
    subject_case: { level: ERROR, rule: never, value: [sentence-case, start-case, pascal-case, upper-case] }
    subject_empty: { level: ERROR, rule: never }
    subject_full_stop: { level: ERROR, rule: never, value: . }
    subject_exclamation_mark: { level: DISABLED, rule: never }
    # Body rules
    body_leading_blank: { level: WARNING, rule: always }
    body_empty: { level: DISABLED, rule: never }
    body_max_line_length: { level: ERROR, rule: always, value: 100 }
    # Footer rules
    footer_leading_blank: { level: WARNING, rule: always }
    footer_empty: { level: DISABLED, rule: never }
    footer_max_line_length: { level: ERROR, rule: always, value: 100 }

# Pull Request Configuration ('pr' command)
pr:
  defaults:
    base_branch: null            # Default base branch (null = repo default)
    feature_prefix: "feature/"   # Default prefix for feature branches

  strategy: github-flow          # Git workflow: github-flow, gitflow, trunk-based

  branch_mapping:                # Branch base/prefix mapping (primarily for GitFlow)
    feature: { base: develop, prefix: "feature/" }
    release: { base: main, prefix: "release/" }
    hotfix: { base: main, prefix: "hotfix/" }
    bugfix: { base: develop, prefix: "bugfix/" }

  generate:                      # Content generation settings
    title_strategy: commits      # How to generate title: commits, llm, template
    description_strategy: commits # How to generate description: commits, llm, template
    # Template used if description_strategy is 'template' AND use_workflow_templates is false
    description_template: |
      ## Changes
      {changes}

      ## Testing
      {testing_instructions}

      ## Screenshots
      {screenshots}
    use_workflow_templates: true # Use built-in templates based on workflow/branch type?

Configuration Priority

The configuration is loaded in the following order (later sources override earlier ones):

1. Default configuration from the package
2. .codemap.yml in the project root
3. Custom config file specified with --config
4. Command-line arguments

Environment Variables

Configuration can also be influenced by environment variables. Create a .env or .env.local file:

# LLM Provider API Keys
OPENAI_API_KEY=your_key_here
ANTHROPIC_API_KEY=your_key_here
GROQ_API_KEY=your_key_here
MISTRAL_API_KEY=your_key_here
COHERE_API_KEY=your_key_here
TOGETHER_API_KEY=your_key_here
OPENROUTER_API_KEY=your_key_here

# Optional: Custom API Base URLs
OPENAI_API_BASE=your_custom_url
ANTHROPIC_API_BASE=your_custom_url

Configuration Tips

1. Token Limits (Deprecated): The token_limit setting is no longer used. Use max_content_length to control file content size.

2. Git Integration

   • gen.use_gitignore: true respects your .gitignore.
   • Configure commit.convention.scopes to match your project structure or leave empty for auto-derivation.
   • Use commit.bypass_hooks or the --bypass-hooks flag to skip Git hooks.

3. LLM Settings

   • Use the global llm section for default model and API base.
   • Override with --model in commit, pr, or ask commands.
   • Set custom API bases (e.g., llm.api_base) for self-hosted or proxy services.

4. Commit Conventions & Linting

   • Customize commit.convention.types and commit.convention.scopes.
   • Adjust commit.convention.max_length.
   • Configure commit.lint rules (based on commitlint standard). The linting section mentioned previously is now under commit.lint.

5. PR Workflow Settings

   • Choose a pr.strategy (github-flow, gitflow, trunk-based).
   • Configure pr.defaults and pr.branch_mapping (especially for GitFlow).
   • Customize PR content generation (pr.generate.title_strategy, pr.generate.description_strategy).
   • Use pr.generate.use_workflow_templates to control template usage. Define a fallback in pr.generate.description_template.

6. Documentation Generation (gen)

   • Control detail level with gen.lod_level or --lod.
   • Enable/disable semantic analysis (gen.semantic_analysis, --semantic).
   • Toggle the directory tree (gen.include_tree, --tree).
   • Configure the Mermaid entity graph (gen.include_entity_graph, gen.mermaid_* options, and corresponding flags).

7. File Watching and Indexing

   • Adjust watcher.debounce_delay to control how quickly the file watcher responds to changes.
   • Use the index command with --watch flag to keep the vector database in sync with code changes.

8. Ask Command Settings

   • Configure ask.interactive_chat to enable interactive mode by default.
   • Override the model with ask.model or use the global llm.model.

Output Structure

The generated documentation includes:

1. Project overview and structure
2. Directory tree visualization
3. Token-optimized code summaries
4. File relationships and dependencies
5. Rich markdown formatting with syntax highlighting

File Processing

The generator:

• Respects .gitignore patterns by default
• Intelligently analyzes code structure
• Optimizes content for token limits
• Generates well-structured markdown
• Handles various file types and languages

Development Setup

Before contributing, please read our Code of Conduct and Contributing Guidelines.

1. Clone the repository:

   git clone https://github.com/SarthakMishra/codemap.git
   cd codemap
   

2. Install Prerequisites:

   • Task: Follow the official installation guide: https://taskfile.dev/installation/
   • uv: Install the uv package manager. We recommend using pipx:

     # Using pipx (recommended)
     pipx install uv
     
     # Or using pip
     # pip install uv
     

   • Python: Ensure you have Python 3.12 or later installed.

3. Set up the Virtual Environment:

   # Create a virtual environment using uv (creates .venv directory)
   uv venv
   
   # Activate the virtual environment
   # On Linux/macOS (bash/zsh):
   source .venv/bin/activate
   # On Windows (Command Prompt):
   # .venv\Scripts\activate.bat
   # On Windows (PowerShell):
   # .venv\Scripts\Activate.ps1
   

4. Install Dependencies: Install project dependencies, including development tools, using uv:

   # Installs dependencies from pyproject.toml including the 'dev' group
   uv sync --dev
   

5. Verify Setup: You can list available development tasks using Task:

   task -l
   

   To run all checks and tests (similar to CI):

   task ci
   

For detailed contribution guidelines, branching strategy, and coding standards, please refer to our Contributing Guide.

Acknowledgments

CodeMap relies on these excellent open-source libraries and models:

Core Dependencies

• LiteLLM (>=1.67.0) - Unified interface for LLM providers
• NumPy (>=2.2.5) - Numerical computing for vector operations
• Pygments (>=2.19.1) - Syntax highlighting for code snippets
• Python-dotenv (>=1.1.0) - Environment variable management
• PyYAML (>=6.0.2) - YAML parsing and configuration management
• Questionary (>=2.1.0) - Interactive user prompts
• Requests (>=2.32.3) - HTTP library for API interactions
• Rich (>=14.0.0) - Beautiful terminal formatting and output
• Typer (>=0.15.2) - Modern CLI framework for Python
• Typing Extensions (>=4.13.2) - Backported typing features
• Sentence-Transformers (>=4.1.0) - Text embeddings for semantic code analysis

Development Tools

• isort (>=6.0.1) - Import sorting
• pylint (>=3.3.6) - Code analysis
• pyright (>=1.1.399) - Static type checking
• pytest (>=8.3.5) - Testing framework
• pytest-cov (>=6.1.1) - Test coverage reporting
• ruff (>=0.11.6) - Fast Python linter

Models

• Code Embeddings: Qodo/Qodo-Embed-1-1.5B - State-of-the-art embedding model optimized for code retrieval tasks.
• LLM Support: Compatible with various providers through LiteLLM including:
  • OpenAI models
  • Anthropic Claude models
  • Groq models
  • Mistral models
  • Cohere models
  • Together AI models
  • OpenRouter providers

Special Thanks

• Cursor
• OpenHands
• GitHub Actions
• Img Shields
• Codecov

License

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