From 941dfe2160ebc05cdd89ad122a1177dbd6c9e1e5 Mon Sep 17 00:00:00 2001
From: Claire <claire.github-309c@sitedethib.com>
Date: Mon, 27 Nov 2023 14:25:54 +0100
Subject: [PATCH] Clamp dates when serializing to Elasticsearch API (#28081)

---
 app/chewy/accounts_index.rb                     |  4 +++-
 app/chewy/concerns/datetime_clamping_concern.rb | 14 ++++++++++++++
 app/chewy/public_statuses_index.rb              |  4 +++-
 app/chewy/statuses_index.rb                     |  4 +++-
 app/chewy/tags_index.rb                         |  4 +++-
 5 files changed, 26 insertions(+), 4 deletions(-)
 create mode 100644 app/chewy/concerns/datetime_clamping_concern.rb

diff --git a/app/chewy/accounts_index.rb b/app/chewy/accounts_index.rb
index 00db257ac711db..59f2f991f2bda0 100644
--- a/app/chewy/accounts_index.rb
+++ b/app/chewy/accounts_index.rb
@@ -1,6 +1,8 @@
 # frozen_string_literal: true
 
 class AccountsIndex < Chewy::Index
+  include DatetimeClampingConcern
+
   settings index: index_preset(refresh_interval: '30s'), analysis: {
     filter: {
       english_stop: {
@@ -60,7 +62,7 @@ class AccountsIndex < Chewy::Index
     field(:following_count, type: 'long')
     field(:followers_count, type: 'long')
     field(:properties, type: 'keyword', value: ->(account) { account.searchable_properties })
-    field(:last_status_at, type: 'date', value: ->(account) { account.last_status_at || account.created_at })
+    field(:last_status_at, type: 'date', value: ->(account) { clamp_date(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: 'verbatim', value: ->(account) { account.searchable_text }) { field :stemmed, type: 'text', analyzer: 'natural' }
diff --git a/app/chewy/concerns/datetime_clamping_concern.rb b/app/chewy/concerns/datetime_clamping_concern.rb
new file mode 100644
index 00000000000000..7f176b6e5489f4
--- /dev/null
+++ b/app/chewy/concerns/datetime_clamping_concern.rb
@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+module DatetimeClampingConcern
+  extend ActiveSupport::Concern
+
+  MIN_ISO8601_DATETIME = '0000-01-01T00:00:00Z'.to_datetime.freeze
+  MAX_ISO8601_DATETIME = '9999-12-31T23:59:59Z'.to_datetime.freeze
+
+  class_methods do
+    def clamp_date(datetime)
+      datetime.clamp(MIN_ISO8601_DATETIME, MAX_ISO8601_DATETIME)
+    end
+  end
+end
diff --git a/app/chewy/public_statuses_index.rb b/app/chewy/public_statuses_index.rb
index b5f0be5e5ce16b..09a4dfc09320a3 100644
--- a/app/chewy/public_statuses_index.rb
+++ b/app/chewy/public_statuses_index.rb
@@ -1,6 +1,8 @@
 # frozen_string_literal: true
 
 class PublicStatusesIndex < Chewy::Index
+  include DatetimeClampingConcern
+
   settings index: index_preset(refresh_interval: '30s', number_of_shards: 5), analysis: {
     filter: {
       english_stop: {
@@ -62,6 +64,6 @@ class PublicStatusesIndex < Chewy::Index
     field(:tags, type: 'text', analyzer: 'hashtag', value: ->(status) { status.tags.map(&:display_name) })
     field(:language, type: 'keyword')
     field(:properties, type: 'keyword', value: ->(status) { status.searchable_properties })
-    field(:created_at, type: 'date')
+    field(:created_at, type: 'date', value: ->(status) { clamp_date(status.created_at) })
   end
 end
diff --git a/app/chewy/statuses_index.rb b/app/chewy/statuses_index.rb
index e315a2030f2363..e739ccecb41b1a 100644
--- a/app/chewy/statuses_index.rb
+++ b/app/chewy/statuses_index.rb
@@ -1,6 +1,8 @@
 # frozen_string_literal: true
 
 class StatusesIndex < Chewy::Index
+  include DatetimeClampingConcern
+
   settings index: index_preset(refresh_interval: '30s', number_of_shards: 5), analysis: {
     filter: {
       english_stop: {
@@ -60,6 +62,6 @@ class StatusesIndex < Chewy::Index
     field(:searchable_by, type: 'long', value: ->(status) { status.searchable_by })
     field(:language, type: 'keyword')
     field(:properties, type: 'keyword', value: ->(status) { status.searchable_properties })
-    field(:created_at, type: 'date')
+    field(:created_at, type: 'date', value: ->(status) { clamp_date(status.created_at) })
   end
 end
diff --git a/app/chewy/tags_index.rb b/app/chewy/tags_index.rb
index 5b6349a964963a..c99218a47fcdcb 100644
--- a/app/chewy/tags_index.rb
+++ b/app/chewy/tags_index.rb
@@ -1,6 +1,8 @@
 # frozen_string_literal: true
 
 class TagsIndex < Chewy::Index
+  include DatetimeClampingConcern
+
   settings index: index_preset(refresh_interval: '30s'), analysis: {
     analyzer: {
       content: {
@@ -42,6 +44,6 @@ class TagsIndex < Chewy::Index
     field(:name, type: 'text', analyzer: 'content', value: :display_name) { field(:edge_ngram, type: 'text', analyzer: 'edge_ngram', search_analyzer: 'content') }
     field(:reviewed, type: 'boolean', value: ->(tag) { tag.reviewed? })
     field(:usage, type: 'long', value: ->(tag, crutches) { tag.history.aggregate(crutches.time_period).accounts })
-    field(:last_status_at, type: 'date', value: ->(tag) { tag.last_status_at || tag.created_at })
+    field(:last_status_at, type: 'date', value: ->(tag) { clamp_date(tag.last_status_at || tag.created_at) })
   end
 end