Add Elasticsearch without external dependencies

[dbackeus]

Why

The goal of the rails-example application is to serve as an example web application for learning how to deploy apps to the Reclaim the Stack platform. Currently it uses Postgres and Redis but since the platform also supports Elasticsearch we want the demo application to use this as well.

What

Indexes the Post and Link models into a searchable Elasticsearch index to allow for doing full text search across both models. Each document in the index will use a title field with a high priority and a content field with lower priority (taken from Post#body and Link#description respectively).

This implementation uses Net::HTTP to interact with Elasticsearch, without involving any extra dependencies.

Feel free to compare with #2 which is based on elasticsearch-rails.

Lines of Code

Implementation: ~200 LOC
Dependencies: 0 LOC

Time spent

Didn't keep track but approximately 2 hours.

Issues

None

Learnings

I learnt a few things about the Elasticsearch API which will come in handy when working with and debugging Elasticsearch in the future. I also learnt how to use Net::HTTP directly, which though not as streamlined as alternative HTTP clients, isn't the end of the world.

[gunnars04]

I would 100% go for without external dependencies. Not everybody is using ruby on rails.