Instead of hardcoding keyword groups into KeywordDB, we know have a
single method for transformi...

- Set up a log handler via the CLI and allow the user to set the log
level
- Switch print stat...

This sorting is done by hand; ultimately it should be handled by a
pre-commit linter.

This sort...

The purpose of this class is to load the YAML keyword database and to
provide an API for sorting...

If a keyword has alternate forms (or synonyms) these can be listed by
turning the keyword into a...

The "keywords" record field is now replaced with fields for the specific
types of keywords:

- a...

This is a first shot at making a machine-readable version of the Learn
Astropy keyword vocabular...

Algolia is case-sensitive when sorting keywords alphabetically in its
facet widget. Thus all cap...

We'll use pre-commit to help validate data files the we have and will be
adding.

See example from https://github.com/algolia/algoliasearch-client-python/pull/404

Tutorial record formatting improvements

This page has the right structure for "Authors", so its useful for
testing the robustness of tha...

This will improve the default display of search results because h1's
typically have no natural c...

These sections don't need to be specifically indexed because they're
already available as metada...

In tutorials, the summary is assumed to be the first paragraph under the
"Summary" heading.

If certain elements can't be found, ReducedTutorial keeps going. We'll
probably want to issue a ...

This field records when the record was indexed.

This fixture reduces a bunch of duplication in effort to get the HTML
content from test pages in...

Convert CI to GitHub Actions

For compatibility with testing sub-dependencies.

This drops Travis CI.

We'll want to modernize some of the testing infrastructure for
consistenc...

Algolia upload tutorials

The idea of workflows is that they're stateless, asynchronous functions
that process a unit of d...

"Records" are types that are made to be turned directly into Algolia
index records. In this firs...

To use it with asyncio, we also install aiohttp and async_timeout.

Reduce tutorial HTML pages

This first reducer will be a proof-of-concept for how reducers can work,
and will reduce tutoria...

We'll use lxml for html parsing.

The idea behind iter_sections is to break a Sphinx-generated HTML pages
into hierarchically-sect...

ReducedTutorial now captures the absolute URLs of any images that are
part of the tutorial conte...

This HTML page is from
http://learn.astropy.org/rst-tutorials/color-excess.html and will be
used...

Basic extraction of title (h1), authors, and keywords metadata from a
tutorial page.

This will contain reducers that convert HTML pages into records that can
be uploaded to Algolia.

Create package

Development dependencies are installed as part of the "dev" extra.

We're going with a fairly modern setuptools-based package setup here
conforming to PEP 518 with ...