feat(autocomplete): Turn off address parts scoring

This change removes several `should` conditions from autocomplete
Elasticsearch queries.

After analysis in #1469 it appears
that these `should` conditions do not particularly influence address
rseults, but hinder POI results from appearing correctly.

While we might want to tweak things in the future, such as applying a
very small scoring boost for matching postal codes, in general for
autocomplete it seems these conditions are overall not helpful.

Closes #1308

query/autocomplete.js

@@ -47,12 +47,6 @@ query.score( views.ngrams_last_token_only_multi( adminFields ), 'must' );
 query.score( views.admin_multi_match_first( adminFields ), 'must');
 query.score( views.admin_multi_match_last( adminFields ), 'must');
 
-// address components
-query.score( peliasQuery.view.address('housenumber') );
-query.score( peliasQuery.view.address('street') );
-query.score( peliasQuery.view.address('cross_street') );
-query.score( peliasQuery.view.address('postcode') );
-
 // scoring boost
 query.score( peliasQuery.view.focus( views.ngrams_strict ) );
 query.score( peliasQuery.view.popularity( peliasQuery.view.leaf.match_all ) );

test/unit/fixture/autocomplete_single_character_street.js

@@ -33,16 +33,6 @@ module.exports = {
         }
       }],
       'should':[
-        {
-          'match': {
-            'address_parts.street': {
-              'query': 'k road',
-              'cutoff_frequency': 0.01,
-              'boost': 1,
-              'analyzer': 'peliasQuery'
-            }
-          }
-        },
         {
         'function_score': {
           'query': {