Records exceed 10 Kb because HTML and content in JSON are the same.

## I want to report a bug: While indexing my blog, the command repeatedly stops because the file to index is too big, mainly caused by the HTML part of the JSON

### What is the current behavior?
When I index my site, I get errors like

The jekyll-algolia plugin detected that one of your records exceeds the 10.00 Kb
record size limit.
```
title:    Bloghelden, 10 jaar webloggen in Nederland
url:      /bloghelden/
size:     11.99 Kb

Most probable keys causing the issue:
   html (5.76 Kb), content (5.74 Kb), tags (0.05 Kb)
```

### What is your expected behavior?
I would expect my relative small blogposts to be indexed. I see from the logfile the index both includes the HTML and the content part. They are almost the same besides the removal of the p-tag in the content part. See enclosed JSON for details.

[jekyll-algolia-record-too-big.log](https://github.com/algolia/jekyll-algolia/files/2726392/jekyll-algolia-record-too-big.log)

### Git repository to reproduce the issue:
https://github.com/frankmeeuwsen/DTD-Blog

### Ruby version used:
ruby 2.4.2p198 (2017-09-14 revision 59899)

### Jekyll version used:
jekyll 3.7.3

Reproduce indexing with the linked DTD-Blog repository and inspect the attached jekyll-algolia-record-too-big.log, focusing on the generated JSON fields and the reported 10.00 Kb limit. Done means the affected small blog posts index successfully without redundant record data causing the size error.

Written by the indexing model from the issue text.