From 7bd5ebb0c50bb481c85ef81816089a04746f762d Mon Sep 17 00:00:00 2001
From: Eugen Rochko <eugen@zeonfederated.com>
Date: Mon, 28 Aug 2023 11:36:17 +0200
Subject: [PATCH] Fix multiple issues with status index mappings (#26686)

---
 app/chewy/accounts_index.rb        |  9 +++++----
 app/chewy/public_statuses_index.rb | 14 ++++++++++----
 app/chewy/statuses_index.rb        | 14 ++++++++++----
 3 files changed, 25 insertions(+), 12 deletions(-)

diff --git a/app/chewy/accounts_index.rb b/app/chewy/accounts_index.rb
index 61e3399aa8..8881b08f66 100644
--- a/app/chewy/accounts_index.rb
+++ b/app/chewy/accounts_index.rb
@@ -21,19 +21,20 @@ class AccountsIndex < Chewy::Index
 
     analyzer: {
       natural: {
-        tokenizer: 'uax_url_email',
+        tokenizer: 'standard',
         filter: %w(
-          english_possessive_stemmer
           lowercase
           asciifolding
           cjk_width
+          elision
+          english_possessive_stemmer
           english_stop
           english_stemmer
         ),
       },
 
       verbatim: {
-        tokenizer: 'standard',
+        tokenizer: 'uax_url_email',
         filter: %w(lowercase asciifolding cjk_width),
       },
 
@@ -62,6 +63,6 @@ class AccountsIndex < Chewy::Index
     field(:last_status_at, type: 'date', value: ->(account) { account.last_status_at || account.created_at })
     field(:display_name, type: 'text', analyzer: 'verbatim') { field :edge_ngram, type: 'text', analyzer: 'edge_ngram', search_analyzer: 'verbatim' }
     field(:username, type: 'text', analyzer: 'verbatim', value: ->(account) { [account.username, account.domain].compact.join('@') }) { field :edge_ngram, type: 'text', analyzer: 'edge_ngram', search_analyzer: 'verbatim' }
-    field(:text, type: 'text', analyzer: 'whitespace', value: ->(account) { account.searchable_text }) { field :stemmed, type: 'text', analyzer: 'natural' }
+    field(:text, type: 'text', analyzer: 'verbatim', value: ->(account) { account.searchable_text }) { field :stemmed, type: 'text', analyzer: 'natural' }
   end
 end
diff --git a/app/chewy/public_statuses_index.rb b/app/chewy/public_statuses_index.rb
index 1fad5de3a1..5c68a13658 100644
--- a/app/chewy/public_statuses_index.rb
+++ b/app/chewy/public_statuses_index.rb
@@ -20,13 +20,19 @@ class PublicStatusesIndex < Chewy::Index
     },
 
     analyzer: {
-      content: {
+      verbatim: {
         tokenizer: 'uax_url_email',
+        filter: %w(lowercase),
+      },
+
+      content: {
+        tokenizer: 'standard',
         filter: %w(
-          english_possessive_stemmer
           lowercase
           asciifolding
           cjk_width
+          elision
+          english_possessive_stemmer
           english_stop
           english_stemmer
         ),
@@ -40,9 +46,9 @@ class PublicStatusesIndex < Chewy::Index
                       .includes(:media_attachments, :preloadable_poll, :preview_cards)
 
   root date_detection: false do
-    field(:id, type: 'keyword')
+    field(:id, type: 'long')
     field(:account_id, type: 'long')
-    field(:text, type: 'text', analyzer: 'whitespace', value: ->(status) { status.searchable_text }) { field(:stemmed, type: 'text', analyzer: 'content') }
+    field(:text, type: 'text', analyzer: 'verbatim', value: ->(status) { status.searchable_text }) { field(:stemmed, type: 'text', analyzer: 'content') }
     field(:language, type: 'keyword')
     field(:properties, type: 'keyword', value: ->(status) { status.searchable_properties })
     field(:created_at, type: 'date')
diff --git a/app/chewy/statuses_index.rb b/app/chewy/statuses_index.rb
index 130f8801df..6d33521051 100644
--- a/app/chewy/statuses_index.rb
+++ b/app/chewy/statuses_index.rb
@@ -20,13 +20,19 @@ class StatusesIndex < Chewy::Index
     },
 
     analyzer: {
-      content: {
+      verbatim: {
         tokenizer: 'uax_url_email',
+        filter: %w(lowercase),
+      },
+
+      content: {
+        tokenizer: 'standard',
         filter: %w(
-          english_possessive_stemmer
           lowercase
           asciifolding
           cjk_width
+          elision
+          english_possessive_stemmer
           english_stop
           english_stemmer
         ),
@@ -64,9 +70,9 @@ class StatusesIndex < Chewy::Index
   end
 
   root date_detection: false do
-    field(:id, type: 'keyword')
+    field(:id, type: 'long')
     field(:account_id, type: 'long')
-    field(:text, type: 'text', analyzer: 'whitespace', value: ->(status) { status.searchable_text }) { field(:stemmed, type: 'text', analyzer: 'content') }
+    field(:text, type: 'text', analyzer: 'verbatim', value: ->(status) { status.searchable_text }) { field(:stemmed, type: 'text', analyzer: 'content') }
     field(:searchable_by, type: 'long', value: ->(status, crutches) { status.searchable_by(crutches) })
     field(:language, type: 'keyword')
     field(:properties, type: 'keyword', value: ->(status) { status.searchable_properties })