From 690919e91b809637de519ae3224f6809595d4ca4 Mon Sep 17 00:00:00 2001
From: Julian Simioni <julian@simioni.org>
Date: Tue, 21 Sep 2021 08:58:48 -0700
Subject: [PATCH] feat(names): Remove variant names from index

Who's on First variant names are a useful collection of unofficial names
for places, but they tend to be pretty messy. This PR explores the
effect of removing them from indexing.

While there might be occasionally useful names in there, it seems like
the majority are exact or near duplicates of more official names, or
names that are so colloquial that they are not particularly useful (do
we _really_ need to support returning NYC for queries for "the big
apple"?).

For reference, here are some variant names for some key places, just to
record the kind of data that's in there:

NYC:
```
Bigapple
NY City
NY Cty
New York City
New York Cty
Newyork
Newyorkcity
Novaiorque
Nycity
Thebigapple
Big Apple
```

San Francisco:
```
S Francisco
S. Francisco
SFO
Sanfran
Sanfrancisco
Frisco
```

China:
```
China - Peoples Republic
China Peoples Rep
China, People's Republic
Chinese
PR China
PR of China
People's Republic of China
Peoples Republic of China
```
---
 src/components/extractFields.js      | 3 +--
 test/components/extractFieldsTest.js | 8 +++-----
 2 files changed, 4 insertions(+), 7 deletions(-)

diff --git a/src/components/extractFields.js b/src/components/extractFields.js
index 7692f735..4865d8b2 100644
--- a/src/components/extractFields.js
+++ b/src/components/extractFields.js
@@ -22,12 +22,11 @@ const population_hierarchy = [
 // note the '%s' is replaced by a language code
 const NAME_ALIAS_FIELDS = [
   'name:%s_x_preferred',
-  'name:%s_x_variant',
   'label:%s_x_preferred_longname',
   'label:%s_x_preferred'
 ];
 
-const WOF_NAMES_REGEX = /^(name|label):[a-z]{3}_x_(preferred|variant)$/;
+const WOF_NAMES_REGEX = /^(name|label):[a-z]{3}_x_preferred$/;
 
 // this function is used to verify that a US county QS altname is available
 function isUsCounty(base_record, wof_country, qs_a2_alt) {
diff --git a/test/components/extractFieldsTest.js b/test/components/extractFieldsTest.js
index 2128f652..3ca6aea1 100644
--- a/test/components/extractFieldsTest.js
+++ b/test/components/extractFieldsTest.js
@@ -948,7 +948,7 @@ tape('name alias tests', (test) => {
       }
     }];
 
-    const expected_name_aliases = ['preferred1', 'preferred2', 'variant1', 'variant2', 'englabel1', 'englabel2'];
+    const expected_name_aliases = ['preferred1', 'preferred2', 'englabel1', 'englabel2'];
 
     test_stream(input, extractFields.create(), function (err, actual) {
       t.deepEqual(actual[0].name_aliases, expected_name_aliases, 'name aliases populated from preferred and variant fields');
@@ -974,7 +974,6 @@ tape('name alias tests', (test) => {
 
     const expected_name_aliases = [
       'preferred1', 'preferred2',
-      'variant1', 'variant2',
       'spalabel1', 'spalabel2',
       'fralabel1', 'fralabel2',
       'englabel1', 'englabel2'
@@ -1003,12 +1002,11 @@ tape('multi-lang index test', (test) => {
 
     const expected_name_langs = {
       'en': ['preferredENG1'],
-      'fr': ['preferredFRA1', 'preferredFRA2'],
-      'es': ['variantSPA1', 'variantSPA2']
+      'fr': ['preferredFRA1', 'preferredFRA2']
     };
 
     test_stream(input, extractFields.create(), function (err, actual) {
-      t.deepEqual(actual[0].name_langs, expected_name_langs, 'name langs populated from fr preferred and SPA variant fields');
+      t.deepEqual(actual[0].name_langs, expected_name_langs, 'name langs populated from fr preferred fields');
       t.end();
     });
   });