A testing framework (plugin + test fixture)
for MkDocs projects

View the documentation on Read the Docs

• A testing framework (plugin + test fixture)for MkDocs projects
  • Description
    • What problem does it solve?
    • MkDocs-Test
  • Usage
    • Installation
      • From pypi
      • Locally (Github)
    • Installing the plugin
    • Performing basic tests
    • Tests on a page
    • Testing the HTML
  • Performing advanced tests
    • Reading the configuration file
    • Accessing page metadata
    • Reading the log
  • License

Description

What problem does it solve?

Traditionally, the quickest way for maintainers of an existing MkDocs website project (or developers of an MkDocs plugin) to check whether an MkDocs project builds correctly, is to run mkdocs build (possibly with the --strict option).

It leaves the following issues open:

• This is a binary proposition: it worked or it didn't.
• It doesn't perform integrity tests on the pages; if something started to go wrong, the issue might emerge only later.
• If something went already wrong, it doesn't necessarily say where, or why.

MkDocs-Test

The purpose of Mkdocs-Test is to facilitate the comparison of source pages (Markdown files) and destination pages (HTML) in an MkDocs project.

MkDocs-Test is a framework composed of two parts:

• an MkDocs plugin (test): it creates a __test__ directory, which contains the data necessary to map the pages of the website.

• a framework for conducting the test. The DocProject class groups together all the information necessary for the tests on a specific MkDocs project.

📝 Technical Note
The plugin exports the nav object, in the form of a dictionary of Page objects.

Usage

Installation

From pypi

pip install mkdocs-test

Locally (Github)

pip install .

Or, to install the test dependencies (for testing this package, not MkDocs projects):

pip install .[test]

Installing the plugin

⚠️ The plugin is a pre-requisite
The framework will not work without the plugin (it exports the pages map into the __test__ directory).

Declare the test plugin in your config file (typically mkdocs.yml):

plugins:
  - search
  - ...
  - test

Performing basic tests

The choice of testing tool is up to you (the examples in this package were tested with pytest).

from mkdocs_test import DocProject

project = DocProject() # declare the project
                       # (by default, the directory where the program resides)
project.build(strict=False, verbose=False)
              # build the project; these are the default values for arguments

assert project.success # return code of the build is zero (success) ?
print(project.build_result.returncode) # return code from the build

# perform automatic integrity checks (do pages exist?)
project.self_check()

Tests on a page

Each page of the MkDocs project can be tested separately

# find the page by relative pathname:
page = project.pages['index.md']

# find the page by name, filename or relative pathname:
page = project.get_page('index')

# easy, and naïve search on the target html page
assert "hello world" in page.html

# find the words "second page", under the header that contains "subtitle"
# at level 2; arguments header and header_level are optional
# the method returns the text so found (if found)
# the search is case insensitive
assert page.find_text_text('second page', header="subtitle", header_level=2)

⚠️ Two markdown versions
page.markdown contains the markdown after possible transformations by plugins; whereas page.source.markdown contains the exact markdown in the file.

If you wish to have the complete source file (including the frontmatter), use page.source.text.

Testing the HTML

You can directly access the .find() and .find_all() methods offered by BeautifulSoup.

page = project.get_page('foo')
headers = page.find_all('h2') # get all headers of level 2
for header in headers:
  print(header.string)

script = page.find('script', type="module")
assert 'import foo' in script.string

Performing advanced tests

Reading the configuration file

print(project.config.site_name)

Accessing page metadata

page = project.get_page('index')
assert page.meta.foo = 'hello' # key-value pair in the page's frontmatter

Reading the log

# search in the trace (naïve)
assert "Cleaning site" in project.trace

# get all WARNING log entries
entries = myproject.find_entries(severity='WARNING')

# get the entry from source 'test', containing 'debug file' (case-insensitive)
entry = project.find_entry('debug file', source='test')
assert entry, "Entry not found"

License

MIT License