Any fool can write code that a computer can understand. Good programmers write code that humans can understand. -- Martin Fowler, 1999
We all love documentation because it makes our codebases easier to understand, yet no one has time to write it in a good and proper way.
Vim-doge is a (Do)cumentation (Ge)nerator which will generate a proper documentation
skeleton based on certain expressions (mainly functions). Simply put your cursor
on a function, press <Leader>d
, jump quickly through TODO
items using
<Tab>
and <S-Tab>
to quickly add descriptions and go on coding!
Every language that has a documentation standard should be supported by vim-doge.
Is your favorite language not supported?
Suggest a new language :tada:
Is your favorite doc standard not supported?
Suggest a new doc standard :tada:
Language | Doc standards | |
---|---|---|
:white_check_mark: | Python | reST, Numpy, Google, Sphinx, Doxygen |
:white_check_mark: | PHP | phpdoc |
:white_check_mark: | JavaScript (Vanilla, ES6, FlowJS, NodeJS, Svelte, Vue) | JSDoc |
:white_check_mark: | TypeScript | JSDoc |
:white_check_mark: | HTML (through inline <script> tags) |
JSDoc |
:white_check_mark: | Lua | LDoc |
:white_check_mark: | Java | JavaDoc |
:white_check_mark: | Groovy | JavaDoc |
:white_check_mark: | Ruby | YARD |
:white_check_mark: | C++ | Doxygen |
:white_check_mark: | C | Doxygen, KernelDoc |
:white_check_mark: | C# | XMLDoc |
:white_check_mark: | Bash | |
:white_check_mark: | Rust | RustDoc |
:white_check_mark: | R | Roxygen2 |
:white_check_mark: | Scala | ScalaDoc |
Using plug:
Plug 'kkoomen/vim-doge', { 'do': { -> doge#install() } }
Using vim-pack:
git clone --recursive --depth 1 https://github.com/kkoomen/vim-doge ~/.vim/pack/*/start/vim-doge
:call doge#install()
Using Pathogen:
git clone --recursive --depth 1 https://github.com/kkoomen/vim-doge ~/.vim/bundle/vim-doge
:call doge#install()
Using Vundle:
Plugin 'kkoomen/vim-doge'
:call doge#install()
Using packer.nvim:
use {
'kkoomen/vim-doge',
run = ':call doge#install()'
}
Using LazyVim
:
{
"kkoomen/vim-doge",
}
:call doge#install()
Run :help doge
to get the full help page.
Vim-doge supports multiple doc standard and you can overwrite them per filetype in your vimrc. Is your favorite doc standard not supported? Suggest a new doc standard :tada:
Example:
let g:doge_doc_standard_python = 'numpy'
If you want to change the doc standard specifically for a buffer, then you can do:
" Inside test.py
:let b:doge_doc_standard = 'numpy'
If you want to generate a docblock using a different doc standard just for a
specific expression, then you can use the DogeGenerate
command:
" Inside test.py, cursor is at a function expression (cursor = `|`):
" |def foo(p1, p2):
" pass
:DogeGenerate numpy
Here is the full list of available doc standards per filetype:
:warning: For C and C++, see C Family Doc Standard Examples.
Variable | Default | Supported |
---|---|---|
g:doge_doc_standard_python |
'reST' |
'reST' , 'numpy' , 'google' , 'sphinx' , 'doxygen' |
g:doge_doc_standard_php |
'phpdoc' |
'phpdoc' |
g:doge_doc_standard_javascript |
'jsdoc' |
'jsdoc' |
g:doge_doc_standard_html |
'jsdoc' |
'jsdoc' |
g:doge_doc_standard_lua |
'ldoc' |
'ldoc' |
g:doge_doc_standard_java |
'javadoc' |
'javadoc' |
g:doge_doc_standard_groovy |
'javadoc' |
'javadoc' |
g:doge_doc_standard_ruby |
'YARD' |
'YARD' |
g:doge_doc_standard_cpp |
'doxygen_javadoc' |
'doxygen_javadoc' , 'doxygen_javadoc_no_asterisk' , 'doxygen_javadoc_banner' , 'doxygen_qt' , 'doxygen_qt_no_asterisk' , 'doxygen_cpp_comment_slash' , 'doxygen_cpp_comment_exclamation' , 'doxygen_cpp_comment_slash_banner' |
g:doge_doc_standard_c |
'doxygen_javadoc' |
'kernel_doc' , 'doxygen_javadoc' , 'doxygen_javadoc_no_asterisk' , 'doxygen_javadoc_banner' , 'doxygen_qt' , 'doxygen_qt_no_asterisk' , 'doxygen_cpp_comment_slash' , 'doxygen_cpp_comment_exclamation' , 'doxygen_cpp_comment_slash_banner' |
g:doge_doc_standard_sh |
'google' |
'google' |
g:doge_doc_standard_rs |
'rustdoc' |
'rustdoc' |
g:doge_doc_standard_cs |
'xmldoc' |
'xmldoc' |
g:doge_doc_standard_r |
'roxygen2' |
'roxygen2' |
g:doge_doc_standard_scala |
'scaladoc' |
'scaladoc' |
g:doge_install_path
Default: /path/to/vim-doge
The path where the bin/ directory will be installed and loaded from. Can be useful if your system is read-only or uses immutable datastructures.
g:doge_enable_mappings
Default: 1
Whether to enable built-in mappings. If you decide to disable this, then you can copy the mappings below and change them to your needs:
" Generate comment for current line.
nmap <silent> <Leader>d <Plug>(doge-generate)
" Interactive mode comment todo-jumping.
nmap <silent> <TAB> <Plug>(doge-comment-jump-forward)
nmap <silent> <S-TAB> <Plug>(doge-comment-jump-backward)
imap <silent> <TAB> <Plug>(doge-comment-jump-forward)
imap <silent> <S-TAB> <Plug>(doge-comment-jump-backward)
smap <silent> <TAB> <Plug>(doge-comment-jump-forward)
smap <silent> <S-TAB> <Plug>(doge-comment-jump-backward)
Or with Lua for Neovim:
-- Generate comment for current line
vim.keymap.set('n', '<Leader>d', '<Plug>(doge-generate)')
-- Interactive mode comment todo-jumping
vim.keymap.set('n', '<TAB>', '<Plug>(doge-comment-jump-forward)')
vim.keymap.set('n', '<S-TAB>', '<Plug>(doge-comment-jump-backward)')
vim.keymap.set('i', '<TAB>', '<Plug>(doge-comment-jump-forward)')
vim.keymap.set('i', '<S-TAB>', '<Plug>(doge-comment-jump-backward)')
vim.keymap.set('x', '<TAB>', '<Plug>(doge-comment-jump-forward)')
vim.keymap.set('x', '<S-TAB>', '<Plug>(doge-comment-jump-backward)')
g:doge_mapping
Default: '<Leader>d'
The mapping to trigger vim-doge. The mapping accepts a count, to select a specific doc standard, if more than one is defined.
g:doge_filetype_aliases
Default:
{
'javascript': [
'javascript.jsx',
'javascriptreact',
'javascript.tsx',
'typescriptreact',
'typescript',
'typescript.tsx',
],
'html': ['svelte', 'vue'],
'java': ['groovy'],
}
Set filetypes as an alias for other filetypes. The key should be the filetype
that is defined in ftplugin/<filetype>.vim
. The value must be a list of 1 or
more filetypes that will be aliases.
Example:
let g:doge_filetype_aliases = {
\ 'javascript': ['typescript']
\}
If you use the above settings and you open myfile.ts
then it will behave like
you're opening a JavaScript filetype.
g:doge_buffer_mappings
Default: 1
Mappings to jump forward/backward are applied as buffer mappings when interactive mode starts and removed when it ends.
g:doge_mapping_comment_jump_forward
Default: '<Tab>'
The mapping to jump forward to the next TODO
item in a comment. Requires
g:doge_comment_interactive
to be enabled.
g:doge_mapping_comment_jump_backward
Default: '<S-Tab>'
The mapping to jump backward to the previous TODO
item in a comment. Requires
g:doge_comment_interactive
to be enabled.
g:doge_comment_interactive
Default: 1
Jumps interactively through all TODO
items in the generated comment.
g:doge_comment_jump_wrap
Default: 1
Continue to cycle among placeholders when reaching the start or end.
g:doge_comment_jump_modes
Default: ['n', 'i', 's']
Defines the modes in which doge will jump forward and backward when interactive
mode is active. For example: removing i
would allow you to use <Tab>
for
autocompletion in insert mode.
:DogeGenerate {doc_standard}
Command to generate documentation. The {doc_standard}
accepts a count or a
string as argument, and it can complete the available doc standards for the
current buffer.
The numeric value should point to an index key from the
b:doge_supported_doc_standards
variable.
The string value should point to a doc standard name listed in the
b:doge_supported_doc_standards
variable.
Below is a list of language-specific configuration and their default values.
JavaScript settings automatically apply for all the default filetypes that are
being aliased, see g:doge_filetype_aliases
.
let g:doge_javascript_settings = {
\ 'destructuring_props': 1,
\ 'omit_redundant_param_types': 0,
\}
destructuring_props
: Whether to generate @param
tags for the destructured
properties in a function expression.
omit_redundant_param_types
: Whether to omit the {type}
part of
parameters and return types when the type is known (i.e. typescript).
let g:doge_php_settings = {
\ 'resolve_fqn': 1
\}
resolve_fqn
: Whether to resolve the FQN based on the use
statements in the
current buffer. The FQN will be resolved for type hints in parameters and the
return type.let g:doge_python_settings = {
\ 'single_quotes': 0,
\ 'omit_redundant_param_types': 1
\}
single_quotes
: Whether to use single quotes for the multiline comments delimiters.
omit_redundant_param_types
: Whether to omit the {type}
part of parameters
and return types when the type is specified in the function itself.
let g:doge_doxygen_settings = {
\ 'char': '@'
\}
char
: Character to use for docblock tags. Use '@'
for @param
or '\'
for \param
.If you're running your vim commands inside a docker, CI or similar environments
with commands such as vim +PlugInstall +qall > /dev/null
then you probably
want to use headless mode. This will not spawn any terminals or progress bars
and will simply run any process by vim-doge in the background.
This feature can be enabled by passing in { 'headless': 1 }
into the
doge#install()
like so: doge#install({ 'headless': 1 })
.
Example using vim-plug:
Plug 'kkoomen/vim-doge', {'do': { -> doge#install({ 'headless': 1 }) }}
To open all the help pages, run :help doge
.
Help or feedback is always appreciated. If you find any bugs, feel free to submit a bug report. If you think vim-doge can be improved, feel free to submit a feature request or a pull request if you have time to help out.
Read the Contribution Guidelines and Code of Conduct when doing contributions.
If you want a new language to be supported but tree-sitter doesn't support it yet, then feel free to create a custom tree-sitter language parser for that language and then we'll integrate it into vim-doge.
Make sure you install Rust and Cargo and you're good to go.
Vim-doge uses a custom command-line interface called
vim-doge-helper, all made in Rust, that is responsible for parsing
code and generating docblocks. Go have a look at the helper
directory if you
want to contribute to the parsers.
If you want to run the tests locally, you should install vader at the same directory level that vim-doge is, so your structure should look like this (using vim-plug example):
~/.vim/plugged/
├── vader.vim
└── vim-doge
After that, you can do the following:
cd /path/to/vim-doge
vim -u test/vimrc
Vader test/**/**/*.vader
to run all tests:bulb: If you're working on specific tests, you can run that specific test only:
Vader test/filetypes/<filetype>/functions.<ext>
.
For more information, see Contribution Guidelines.
I created vim-doge mainly because I couldn't find a plugin that could generate proper comments for a big collection of languages that I would use on a daily basis in a quick and easy way.
Rather than scraping off the internet to find all sorts of vim plugins for every language I was coding in, I was more or less looking for a single plugin that supports every language I was working in.
Another big motivation for me is that I've noticed people tend to skip the documentation part because writing just the skeleton of the comment takes already too much time and I am one of those people. Having the skeleton generated and an interactive mode to quickly add descriptions is a big time saver.
Do you enjoy using vim-doge? Give it a star on GitHub and submit your vote on vim.org.
Vim-doge generates docblocks, but perhaps you want to take it a step further and generate it into HTML. Dooku.nvim is another plugin that can suit these needs.
Vim-doge is licensed under the GPL-3.0 license.