Mkdocs-caption

* Fix bug in the post processing which cause references with ids which contain other ids
to be replaced with the wrong text.
* Adapt figure generation to allow packing multiple images into the same figure.
This is necessary to e.g allow having different images for light and dark mode.
* Fix bug in the post processing which causes the windows build to fail.
* Fix bug in the post processing which causes incorrect substitutions when a
reference to a target is placed before the target.

* Allow automatic link text generation across different pages.

* Allow markdown syntax within the caption element.

* Prevent conversion of emojis into figure elements with captions.
* Add new config parameter `ignore_classes` to explicitly prevent conversion
of specific images into figures.

* Prevent conversion of inline images even if they have an alt text (9)

* Allow caption elements within indented blocks (e.g. admonitions). Can be disabled
with the `allow_indented_caption` option.