Remove the fallback to parsing the HTML when `rtd_search_projects` is not set

[Cadair]

Provide a general description of the issue or problem.

This section of the code:

sunpy-sphinx-theme/src/sunpy_sphinx_theme/theme/sunpy/static/js/rtd_enhanced_search.js

Lines 80 to 92 in 3961b64

	} else {
	// Use the documentation nav links
	document
	.getElementById("Documentation")
	.querySelectorAll(".nav-link")
	.forEach(function (el) {
	// We might want to somehow limit which projects to include in the search
	let slug = el.innerHTML.replace(/(^\s|\s$)/, "");
	let link = el.getAttribute("href");
	_obj.projects[slug] = { name: slug, url: link };
	});
	this.projectorder = Object.keys(this.projects);
	}

Isn't really needed now we have the Python theme configuration.

@slowe How would you recommend we remove this? I think we would need to just fallback to the original sphinx search? I can make the Python configuration raise an error, so it should be unlikely to hit this.

[slowe]

The enhanced search builds on top of the existing search box so the simplest way to stop it is just to return at this point i.e.:

} else {
    console.warn('Falling back to built-in search');
    return this;
}

Obviously you don't need to include that console.warn() - I've just put that there in case you want to see a message in the console when it falls back.

This way the existing search should be unaffected.