From 91db45b197377750a88c0ac236554451dc6b567d Mon Sep 17 00:00:00 2001
From: Eugen Rochko <eugen@zeonfederated.com>
Date: Fri, 25 Apr 2025 12:35:21 +0200
Subject: [PATCH] Change account search to be more forgiving of spaces (#34455)

---
 app/chewy/accounts_index.rb            | 16 ++++++++++++
 app/services/account_search_service.rb | 36 ++++++++++++++++++++++----
 2 files changed, 47 insertions(+), 5 deletions(-)

diff --git a/app/chewy/accounts_index.rb b/app/chewy/accounts_index.rb
index 59f2f991f2..796584e9c6 100644
--- a/app/chewy/accounts_index.rb
+++ b/app/chewy/accounts_index.rb
@@ -19,9 +19,16 @@ class AccountsIndex < Chewy::Index
         type: 'stemmer',
         language: 'possessive_english',
       },
+
+      word_joiner: {
+        type: 'shingle',
+        output_unigrams: true,
+        token_separator: '',
+      },
     },
 
     analyzer: {
+      # "The FOOING's bar" becomes "foo bar"
       natural: {
         tokenizer: 'standard',
         filter: %w(
@@ -35,11 +42,20 @@ class AccountsIndex < Chewy::Index
         ),
       },
 
+      # "FOO bar" becomes "foo bar"
       verbatim: {
         tokenizer: 'standard',
         filter: %w(lowercase asciifolding cjk_width),
       },
 
+      # "Foo bar" becomes "foo bar foobar"
+      word_join_analyzer: {
+        type: 'custom',
+        tokenizer: 'standard',
+        filter: %w(lowercase asciifolding cjk_width word_joiner),
+      },
+
+      # "Foo bar" becomes "f fo foo b ba bar"
       edge_ngram: {
         tokenizer: 'edge_ngram',
         filter: %w(lowercase asciifolding cjk_width),
diff --git a/app/services/account_search_service.rb b/app/services/account_search_service.rb
index dab5f748bf..5261040884 100644
--- a/app/services/account_search_service.rb
+++ b/app/services/account_search_service.rb
@@ -128,11 +128,37 @@ class AccountSearchService < BaseService
 
     def core_query
       {
-        multi_match: {
-          query: @query,
-          type: 'best_fields',
-          fields: %w(username^2 display_name^2 text text.*),
-          operator: 'and',
+        dis_max: {
+          queries: [
+            {
+              match: {
+                username: {
+                  query: @query,
+                  analyzer: 'word_join_analyzer',
+                },
+              },
+            },
+
+            {
+              match: {
+                display_name: {
+                  query: @query,
+                  analyzer: 'word_join_analyzer',
+                },
+              },
+            },
+
+            {
+              multi_match: {
+                query: @query,
+                type: 'best_fields',
+                fields: %w(text text.*),
+                operator: 'and',
+              },
+            },
+          ],
+
+          tie_breaker: 0.5,
         },
       }
     end