restrict DocSearch to relevant parts of the site

[orbeckst]

The new AlgoliaDoc Search from #73 indexes a large part of our site, as shown in the config file https://github.com/algolia/docsearch-configs/blob/master/configs/mdanalysis.json that was generated by the fine folks at Algolia. They already sensibly excluded code directories. But we still get hits that look like duplicates

• release docs https://www.mdanalysis.org/docs/ and development docs https://www.mdanalysis.org/mdanalysis
• blog posts (eg https://www.mdanalysis.org/2017/06/04/gsoc-students/) and paginated blog posts (eg https://www.mdanalysis.org/blog/page3/)

I would like to restrict the search to make the hits more relevant

• only show hits from the docs for releases
• only show blog posts themselves, not paginated versions
• use sitemap.xml, which is generated with jekyll-sitemap plugin

  MDAnalysis.github.io/_config.yml

  Line 31 in 549bfb4

  - jekyll-sitemap

  in order to include all pages, especially any blog pages that might not show up in pagination (now that we omit pagination)

Unless anyone in @MDAnalysis/coredevs has any objections I will submit a PR to https://github.com/algolia/docsearch-configs to change the index paths.

[orbeckst]

Following the docs ... https://community.algolia.com/docsearch/documentation/docsearch-scraper/config-options/

[orbeckst]

Hm, did not realize that the comment in the other commit would close this one...

[orbeckst]

Closed with algolia/docsearch-configs#409