pytest-check-links 0.5.2

Check links in files

pytest-check-links

pytest plugin that checks URLs for HTML-containing files.

Supported files

• .html
• .rst
• .md (TODO: select renderer)
• .ipynb (requires nbconvert)

Install

pip install pytest-check-links

Use

pytest --check-links mynotebook.ipynb

Configure

--links-ext

default: md,rst,html,ipynb

A comma-separated list of extensions to check

--check-anchors

Also check whether links with #anchors HTML files (either local, or with served with html in the Content-Type) actually exist, and point to exactly one named anchor.

--check-links-ignore

A regular expression that matches URIs that should not be checked. Can be specified multiple times for multiple ignore patterns. This can be used for files that have a lot of links to GitHub pages, such as a Changelog. GitHub has rate limiting, which would normally cause these files to take up to an hour to complete for larger repositories. For example:

pytest --check-links --check-links-ignore "https://github.com/.*/pull/.*" CHANGELOG.md

Cache

Caching requires the installation of requests-cache.

pip install requests-cache

If enabled, each occurance of a link will be checked, no matter how many times it appears in a collection of files to check.

--check-links-cache

Cache requests when checking links. Caching is disabled by default, and this option must be provided, even if other cache configuration options are provided.

--check-links-cache-name

default: .pytest-check-links-cache

Name of link cache, either the base name of a file or similar, depending on backend.

--check-links-cache-backend

default: sqlite3

Cache persistence backend. The other known backends are:

• memory
• redis
• mongodb

See the requests-cache documentation for more information.

--check-links-cache-expire-after

default: None (unlimited)

Time to cache link responses (seconds).

--check-links-cache-backend-opt

Backend-specific options for link cache, provided as key:value. These are passed directly to the requests_cache.CachedSession constructor, as they vary depending on the backend.

Values will be parsed as JSON first, so to overload the default of caching all HTTP response codes (which requires a list of ints):

--check-links-backend-opt allowable_codes:[200]

TODO

• pick a markdown renderer (probably commonmark) or make the markdown renderer pluggable
• options for validating links (allow absolute links, only remote or local, etc.)
• find URLs in Python docstrings