Bump actions/setup-python from 3 to 4

Bumps [actions/setup-python](https://github.com/actions/setup-python) from 3 to 4.
- [Release no...

Bumps [pre-commit/action](https://github.com/pre-commit/action) from 2.0.3 to 3.0.0.
- [Release ...

Update for new nbconvert html format

The latest nbconvert HTML template (used by nbcollection) no longer makes
.jp-RenderedHTMLCommon...

Update pre-commit hooks

In particular, the black update is needed to resolve an incompatibility
with click.

Bump actions/cache from 2 to 3

Bumps [actions/cache](https://github.com/actions/cache) from 2 to 3.
- [Release notes](https://g...

Bump actions/setup-python from 2 to 3

Bump actions/checkout from 2 to 3

Bumps [actions/setup-python](https://github.com/actions/setup-python) from 2 to 3.
- [Release no...

Bumps [actions/checkout](https://github.com/actions/checkout) from 2 to 3.
- [Release notes](htt...

Fixes and improvements for tutorial indexing

The maximum record size in the free plan is 10KB, so we need to split
records that are larger th...

The previous reducer was doing extra recursion of content elements,
resulting in bloated Algolia...

This was accidentally left as a typing-only import, whereas the
index_tutorial_from_path workflo...

Add ability to index tutorials from local files

This allows us to run a single command in a CI environment that points
to the rendered tutorials...

Passing --path allows the index tutorial command to get HTML from a
local build, rather than sou...

This provides a built-in way of creating the HtmlPage resource from
locally-available data.

Ref...

This splits the index_tutorial workflow into an index_tutorial core
workflow that takes a pre-do...

Index nbcollection-built tutorial pages

This is an nbcollection-generated version of the Coordinates-Transforms
tutorial. Note that this...

Different notebook-to-html converters will have different HTML
structures, so it makes sense to ...

Add Priority field

Adjust "importance" field for Guides

Before a user starts to enter a search term on the Learn Astropy
homepage, we want the search re...

This is now possible because JupyterBookMetadata contains the homepage's
URL.

The first record of the guides homepage now exclusively has an
importance of 1; the first record...

This will allow us to know if a page is actually the homepage, and thus
should be treated with a...

Initial JupyterBook ingest feature

The mock hasn't been fully developed so it currently doesn't count
network requests. We'll docum...

Co-authored-by: Adrian Price-Whelan <[email protected]>

Co-authored-by: Adrian Price-Whelan <[email protected]>

New tutorials now use <section> tags around each content chunk, rather
than <div> tags with a cl...

Previously we were also accidentally selecting from the on-page nav. Now
by explicitly selecting...

This is the same content as the original color excess tutorial, but the
HTML structure is now di...

At the end of an indexing workflow, now run the expire_old_records
workflow that finds records f...

The index_epoch field in the Algolia record schemas is for associating a
unique identifier for a...

The issue is that algoliasearch's save_objects() wants python objects
rather than JSON-encoded s...

This command enables us to delete all records associated with a root_url
(useful for debugging).

The logging at "info" level (-v) provides summaries of what URLs are
indexed and how many record...

This is a decision we might want to re-address, but for this this commit
effectively omits Jupyt...

This context manager does a few things for us:

1. Rather than needing to pass around both the S...

The HtmlComment type is generating a ValueError from .text_content().
For now, handle this excep...

The precise name "objectID" is required by Algolia.

Includes both "index tutorial" and "index guide" subcommands.

The new CLI replaces this argparse-based command and will incorporate
the functionality of inges...

Add this static method to the base AlgoliaRecord so that any record
being built from a Section c...

- Use typer framework, which is adds annotations to click
- Design the CLI are commands. There's...

The index_jupyterbook workflow identifies pages in a JupyterBook, and
for each page, launches an...

Similar to TutorialRecord, the GuideRecord is a pydantic model for an
Algolia record of a piece ...

Since Jupyter Book is based on Sphinx, we can use the same
iter_sphinx_sections function as we u...

This field is the overall summary for a documentation project (a
tutorial or a guide). Each reco...

And move to astropylibrarian.reducers. This dataclass now becomes the
basis for reducing content...

This replaces TutorialSectionRecord with the pydantic model,
TutorialRecord.

This change allows...

I'm having trouble with aliases in sub-models. For now I'm dropping
aliases, but may bring them ...

AlgoliaRecord will serve as a base model that individual content types
can subclass to provide a...

This metadata provides context while indexing individual pages from a
JupyterBook.

New class Ju...

The new HtmlPage class provides context and utilities around a
downloaded HTML page, such as the...

The key problem this workflow is tackling at first is that the
"homepage" of a Jupyter Book is n...

Get the URLs of all pages in a Jupyter Book from the 'nav' element.

This test dataset is from [A guide to CCD data reduction and stellar
photometry using astropy an...

Format code with Black and isort

Black formats code and isort formats imports. Both run via pre-commit.
Flake8 runs after these t...

Update testing around tox

This splits the CI up so that linting and type checking are separate
from running unit tests; un...

This will help us keep up to date with our dependencies and their bug
fixes a bit better.

This handles some intricacies of the pre-commit and tox set up: namely
that pre-commit hooks nee...

This is my go-to configuration.

The tests were missing return types and also types for the pytest
fixture.

This also adds flake8 to the "linting" tox environment, from the unit
test environment. This is ...

This lint environment is a front-end for running pre-commit on all
files.

Tox will let us improve our testing and CI workflow.

Sort tutorial keywords into types with a keyword vocabulary