Fixed search results were not limited to the selected language.

[fabianmarz]

By default the index language is set to english and uses the stop word filter which contains the word it. Since this string is also the locale for the Italian language posts were not limited or not showing up when searching in this particular language. This could be fixed by changing the search field from an analysed text field to the keyword field.

Query:

GET /<my_index>/_analyze
{
  "field": "post_lang", 
  "text" : "it"
}

Result:

{
  "tokens" : [ ]
}

---

Query:

GET /<my_index>/_analyze
{
  "field": "post_lang.keyword", 
  "text" : "it"
}

Result:

{
  "tokens" : [
    {
      "token" : "it",
      "start_offset" : 0,
      "end_offset" : 2,
      "type" : "word",
      "position" : 0
    }
  ]
}

---

When not customized, the filter removes the following English stop words by default:
a, an, and, are, as, at, be, but, by, for, if, in, into, is, it, no, not, of, on, or, such, that, the, their, then, there, these, they, this, to, was, will, with
Source: https://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-stop-tokenfilter.html

More information:

• https://stackoverflow.com/a/48875105/2862127
• https://www.elastic.co/blog/strings-are-dead-long-live-strings

[strategio]

Hi @fabianmarz, and thank you for your contribution.

We are collecting issues related to the WPML ElasticPress plugin before to schedule another dev cycle (we have no schedule yet for this).

Be sure that we'll take your suggestion into consideration (even if not in the short-term). Meanwhile, I guess you can live with your patch.

Thanks again!

Internal Ref: wpmlbridge-263

[decodekult]

This is covered now by #13 so no need to enforce the keyword usage.

Thanks @fabianmarz for bringing this up. I am closing this MR as obsolete.