hatch-mkdocs

hatch-mkdocs

Hatch plugin to integrate MkDocs and infer dependencies into an env

This plugin populates Hatch environments with dependencies on the fly based on a mkdocs.yml file.

This is intended to effortlessly manage dependencies for a MkDocs site.

You just need to add this minimal config to Hatch, along with any existing MkDocs config:

hatch.toml	mkdocs.yml
[env] requires = [ "hatch-mkdocs", ] [env.collectors.mkdocs.docs] path = "mkdocs.yml"	site_name: MkDocs example plugins: - autorefs markdown_extensions: - callouts - pymdownx.superfences

This gets you the following implied Hatch configuration: (click to expand)

hatch.toml
[envs.docs] detached = true dependencies = [ "markdown-callouts", "mkdocs", "mkdocs-autorefs", "pymdown-extensions", ] [envs.docs.scripts] build = "mkdocs build -f mkdocs.yml {args}" serve = "mkdocs serve -f mkdocs.yml {args}" gh-deploy = "mkdocs gh-deploy -f mkdocs.yml {args}"

(this is just for posterity, no such config is actually written to a file)

With this:

• You don't need to specify the PyPI dependencies, they get inferred on the fly just from mkdocs.yml by doing a reverse lookup of MkDocs plugins in the catalog, using mkdocs get-deps. (See more details there)

• An automatically managed virtual environment with pre-defined MkDocs commands is at your fingertips.

You can check this yourself:

hatch env show docs

                    Standalone                     
┌──────┬─────────┬────────────────────┬───────────┐
│ Name │ Type    │ Dependencies       │ Scripts   │
├──────┼─────────┼────────────────────┼───────────┤
│ docs │ virtual │ markdown-callouts  │ build     │
│      │         │ mkdocs             │ gh-deploy │
│      │         │ mkdocs-autorefs    │ serve     │
│      │         │ pymdown-extensions │           │
└──────┴─────────┴────────────────────┴───────────┘

The dependencies get resolved and installed into a virtual environment as part of a Hatch invocation. So, you can directly run:

hatch run docs:build

Creating environment: docs
Checking dependencies
Syncing dependencies
INFO    -  Cleaning site directory
INFO    -  Building documentation to directory: site
INFO    -  Documentation built in 0.03 seconds

(If you've been using virtualenvs directly, this single command replaces creating an environment, installing dependencies into it, as well as running mkdocs in it, optionally with arguments)

Furthermore, whenever the set of dependencies changes (i.e. you select new MkDocs plugins), these Hatch commands will re-install dependencies as necessary.
Otherwise, the environment is just reused; the installation happens only on the first invocation.

If at any point you want to make sure the dependencies are re-installed anew, you can just remove the environment:

hatch env remove docs

Removing environment: docs

Installation

Just install Hatch. Ideally in an isolated way with pipx install hatch, or just pip install hatch as a more well-known way.

If you declare hatch-mkdocs as a dependency in your Hatch config (pyproject.toml or hatch.toml) as shown above, Hatch will automatically install it on first use.

Alternatively you can install it manually: pipx inject hatch hatch-mkdocs or just pip install hatch-mkdocs.

And do not install MkDocs - it's unnecessary, only the sub-environments will have it.

Configuration

Note that although Hatch is typically associated with managing entire Python projects and applications, you can use it purely for environment management for a MkDocs site - through this plugin, or even without it.

Hatch can be configured through one of two files - hatch.toml or pyproject.toml. Configs in the latter are equivalent but will always need a [tool.hatch...] prefix; it can be used if you have an existing Python project and you don't want to add another config file.

So, add the following into one of the files:

hatch.toml	pyproject.toml
[env] requires = [ "hatch-mkdocs", ] [env.collectors.mkdocs.ENV_NAME] path = "path/to/mkdocs.yml" [envs.ENV_NAME] ...	[tool.hatch.env] requires = [ "hatch-mkdocs" ] [tool.hatch.env.collectors.mkdocs.ENV_NAME] path = "path/to/mkdocs.yml" [tool.hatch.envs.ENV_NAME] ...

Here, [env.collectors.mkdocs.ENV_NAME] means: please populate an environment named "ENV_NAME" based on an MkDocs config. In that section, path is the path to mkdocs.yml.

At the moment that is the entire configurability of this plugin.

In the first example we used "docs" as the environment name, you can use "mkdocs" as well if you like, or anything else. Further, if you use "default" as the name (which you might do if documentation building is all that you'll ever use Hatch for) then you can skip the environment prefix (docs: in the above example).

Multiple separate environments with their own configs and dependencies can be populated as well.

Inside [envs.ENV_NAME] (which is an ordinary construct in Hatch) you can proceed to further customize the environment (though normally it shouldn't be necessary, and the section can be omitted from the text config): you can add extra dependencies or scripts, or any other environment config. You could also set detached back to false if the documentation actually relies on the project itself being installed, such as in the case of mkdocstrings.