Wingman: Your AI-Powered Coding Companion for Neovim 🚀

Welcome to Wingman, the ultimate plugin that supercharges your Neovim experience by integrating a powerful LLM (Large Language Model) directly into your coding workflow! Imagine having an intelligent assistant that understands your project context and helps you write better code, all while you focus on what you do best—creating amazing software!

Why Wingman?

In today's fast-paced development environment, efficiency is key. Wingman takes your coding to the next level by automatically injecting relevant project context into requests sent to the LLM. This means you get smarter, more contextual responses tailored to your specific needs, making your coding sessions more productive and enjoyable.

Key Features

• Context-Aware Requests: Wingman automatically collects and incorporates project context into your requests, ensuring the LLM fully understands your current work.
• Seamless Integration: Specifically designed for Neovim, Wingman integrates smoothly into your existing workflow, enhancing your coding experience without any hassle.
• User-Friendly Interface: Featuring a clean and intuitive interface, interacting with the LLM is as simple as pressing a few keys. You'll be amazed at how quickly you can get the assistance you need!
• Code Replacement (WIP): Wingman aims to automatically update affected lines of code based on the LLM's suggestions, enabling you to implement improvements with just a few keystrokes.
• Extend Models / Providers Support (WIP): Currently, Wingman supports only OpenAI's models, but new models and providers will be added soon. Contributions are welcome.
• Persist Chat Sessions (WIP): Currently, each time you open the prompt dialog, the conversation's scope is limited to the current buffer context. There are plans to optionally allow session persistence across different dialogs.

Getting Started

To get started with Wingman, simply install the plugin using your favorite package manager. Once installed, you can begin leveraging the power of LLMs in your coding routine.

Installation with LazyVim

1. Add the following to ~/.config/nvim/lua/plugins/wingman.lua configuration:

return {
 "webpolis/wingman.nvim",
 dependencies = {'nvim-lua/plenary.nvim', 'MunifTanjim/nui.nvim', 'junegunn/fzf.vim', 'kkharji/sqlite.lua', 'leafo/lua-openai'},
 opts = {
  openai_api_key = 'your_openai_api_key',
  openai_model = 'gpt-4o-mini'
 }
}

Installation with Packer

1. Add the following to your packer configuration:

return require('packer').startup(function(use)
  -- Dependencies
  use 'nvim-lua/plenary.nvim'
  use 'MunifTanjim/nui.nvim'
  use 'junegunn/fzf.vim'
  use 'kkharji/sqlite.lua'
  use 'leafo/lua-openai'

  use {
    'webpolis/wingman.nvim',
    config = function()
      require('wingman').setup {
        openai_api_key = 'your_openai_api_key',
        openai_model = 'gpt-4o-mini'
      }
    end
  }
end)

Issues with Luarocks / lua-openai

To stream the model's response, certain Lua packages are necessary. If the installation of leafo/lua-openai fails, install it separately using Luarocks and then restart the editor:

luarocks install lua-openai

Configuration Options

Option	Description
openai_api_key	Your OpenAI API key for authentication.
openai_model	The model to use for generating responses.

Usage

1. Open a file in Neovim.
2. Run the :Wingman command.
3. Ask questions or request code suggestions, and watch as Wingman enhances your coding experience!

When the initial prompt dialog opens, Wingman automatically gathers all symbols present in the current buffer. As you type, suggestions will appear as markdown links (e.g. [MySymbol](src/symbols.ts)). Once a suggestion is selected, it will be included in the current prompt's context.

To ensure symbols are extracted, your editor's LSP must be correctly configured. Use LspInfo to verify that the current language is supported by an LSP server. Otherwise, the entire contents of the buffer will be included.

Command	Description
Wingman	Opens the prompt dialog
WingmanCollect	Adds the current buffer's content to the Wingman context without opening a dialog
WingmanClear	Clears the cache and resets the database

Contribute

As Wingman is still a work in progress, your feedback and contributions are invaluable! Join our community, report issues, and help us shape the future of this plugin.