This is a collection of configuration files for the tools that I like to use during software development: Neovim, Tmux (not so much anymore due to Neovide and my homegrown session management), Alacritty.
This project uses GNU Stow to create symlinks from this repository to your $HOME directory.
NOTE:
- This script will ask for sudo permissions, which are then used for brew and apt (if applicable) installs
- This script will create symlinks to ~/.zshrc, ~/.config/nvim and ~/.config/tmux, see GNU Stow for how conflicts will be handled
git clone https://github.com/WizardStark/dotfiles.git
cd dotfiles
./setup.shThe best sources to consult for understanding the nvim config:
- Lazy.nvim - Plugin manager
- Legendary.nvim - Command palette
The most important keybinds are <space><space> in Neovim for the command palette,
wherein you can fuzzy find your way through most available commands, and <C-a>? for
a list of tmux binds - this is much less nice to use as I have not found a way to add
descriptions, but the commands are pretty self explanatory.
I have a venv wrapper script that places all venvs at ~/.virtualenvs.
Relevant commands are lsvenv, mkvenv and rmvenv - they do what you think, and have autocomplete.
Then to activate a venv just do venv my_venv
For full nvim compatibility, I would recommend the following:
mkvenv nvim
venv nvim
pip install pynvim jupyter_clientAnd add the following to ~/.config/nvim/lua/lcl/options.lua
vim.g.python3_host_prog = vim.fn.expand("~/.virtualenvs/nvim/bin/python3")Should you then also want to run jupyter notebooks in vim, for each project do the following (I will probably write a ui wrapper in nvim for this at some point):
venv project_name # activate the project venv
pip install ipykernel
python -m ipykernel install --user --name project_nameRecent versions of jupyter_client also do not create their runtime directory for some reason, so
if you see an error to the effect of "file/directory does not exist /some/path/Jupyter/runtime/kernel-someid",
simply create the directory.
Install Ollama:
curl -fsSL https://ollama.com/install.sh | sh
And currently it appears that qwen2.5-coder is the best local model, so choose from https://ollama.com/library/qwen2.5-coder:7b-instruct (it has been prefilled with a decent choice for 8GB vram gpu's)