From bf34294fcb4e184a5af9fb6f18d663d647d8557a Mon Sep 17 00:00:00 2001
From: Vyr Cossont <VyrCossont@users.noreply.github.com>
Date: Fri, 30 Dec 2022 15:11:27 -0800
Subject: [PATCH] Add is:, has:, domain:, lang: search filters

---
 app/chewy/statuses_index.rb             |  4 ++++
 app/lib/search_query_parser.rb          |  2 +-
 app/lib/search_query_transformer.rb     | 21 ++++++++++++++++++---
 app/models/status.rb                    | 20 ++++++++++++++++++++
 app/services/fetch_link_card_service.rb |  8 ++++++++
 5 files changed, 51 insertions(+), 4 deletions(-)

diff --git a/app/chewy/statuses_index.rb b/app/chewy/statuses_index.rb
index e6f9a04c30..17df16b612 100644
--- a/app/chewy/statuses_index.rb
+++ b/app/chewy/statuses_index.rb
@@ -67,6 +67,10 @@ class StatusesIndex < Chewy::Index
     field :account_id, type: 'long'
     field :created_at, type: 'date'
     field :visibility, type: 'keyword'
+    field :domain, type: 'keyword', value: ->(status) { status.account.domain or Rails.configuration.x.local_domain }
+    field :lang, type: 'keyword', value: ->(status) { status.language }
+    field :is, type: 'keyword', value: ->(status) { status.searchable_is }
+    field :has, type: 'keyword', value: ->(status) { status.searchable_has }
 
     field :text, type: 'text', value: ->(status) { status.searchable_text } do
       field :stemmed, type: 'text', analyzer: 'content'
diff --git a/app/lib/search_query_parser.rb b/app/lib/search_query_parser.rb
index 15956d4cfd..5d6ffbf29d 100644
--- a/app/lib/search_query_parser.rb
+++ b/app/lib/search_query_parser.rb
@@ -9,7 +9,7 @@ class SearchQueryParser < Parslet::Parser
   rule(:prefix)    { (term >> colon).as(:prefix) }
   rule(:shortcode) { (colon >> term >> colon.maybe).as(:shortcode) }
   rule(:phrase)    { (quote >> (term >> space.maybe).repeat >> quote).as(:phrase) }
-  rule(:clause)    { (prefix.maybe >> operator.maybe >> (phrase | term | shortcode)).as(:clause) }
+  rule(:clause)    { (operator.maybe >> prefix.maybe >> (phrase | term | shortcode)).as(:clause) }
   rule(:query)     { (clause >> space.maybe).repeat.as(:query) }
   root(:query)
 end
diff --git a/app/lib/search_query_transformer.rb b/app/lib/search_query_transformer.rb
index d1671459c1..4cba60aaea 100644
--- a/app/lib/search_query_transformer.rb
+++ b/app/lib/search_query_transformer.rb
@@ -35,6 +35,8 @@ class SearchQueryTransformer < Parslet::Transform
         { multi_match: { type: 'most_fields', query: clause.term, fields: ['text', 'text.stemmed'] } }
       when PhraseClause
         { match_phrase: { text: { query: clause.phrase } } }
+      when PrefixClause
+        { term: { clause.filter => clause.term } }
       else
         raise "Unexpected clause type: #{clause}"
       end
@@ -99,10 +101,21 @@ class SearchQueryTransformer < Parslet::Transform
   class PrefixClause
     attr_reader :filter, :operator, :term, :order
 
-    def initialize(prefix, term)
+    def initialize(prefix, operator, term)
+      case operator
+      when '+', nil
+        @operator = :filter
+      when '-'
+        @operator = :must_not
+      else
+        raise "Unknown operator: #{str}"
+      end
+
       case prefix
+      when 'domain', 'is', 'has', 'lang'
+        @filter = prefix.to_s
+        @term = term
       when 'from'
-        @operator = :filter
         @filter = :account_id
 
         username, domain = term.gsub(/\A@/, '').split('@')
@@ -111,6 +124,8 @@ class SearchQueryTransformer < Parslet::Transform
 
         @term = account.id
       when 'sort'
+        raise Mastodon::SyntaxError unless operator.nil?
+
         @operator = :order
         @term = :created_at
 
@@ -133,7 +148,7 @@ class SearchQueryTransformer < Parslet::Transform
     operator = clause[:operator]&.to_s
 
     if clause[:prefix]
-      PrefixClause.new(prefix, clause[:term].to_s)
+      PrefixClause.new(prefix, operator, clause[:term].to_s)
     elsif clause[:term]
       TermClause.new(prefix, operator, clause[:term].to_s)
     elsif clause[:shortcode]
diff --git a/app/models/status.rb b/app/models/status.rb
index fa1c95166f..4b5dcc5111 100644
--- a/app/models/status.rb
+++ b/app/models/status.rb
@@ -173,6 +173,26 @@ class Status < ApplicationRecord
     ].compact.join("\n\n")
   end
 
+  def searchable_is
+    keywords = []
+    keywords << :bot if account.bot?
+    keywords << :local if local?
+    keywords << :local_only if local_only
+    keywords << :reply if reply?
+    keywords
+  end
+
+  def searchable_has
+    keywords = []
+    keywords << :cw if spoiler_text?
+    keywords << :link if FetchLinkCardService.new.link?(self)
+    keywords << :media if media_attachments.present?
+    keywords << :mention if mentions.present?
+    keywords << :poll if preloadable_poll.present?
+    keywords << :tag if tags.present?
+    keywords
+  end
+
   def to_log_human_identifier
     account.acct
   end
diff --git a/app/services/fetch_link_card_service.rb b/app/services/fetch_link_card_service.rb
index e5b5b730ec..86fe4d96b9 100644
--- a/app/services/fetch_link_card_service.rb
+++ b/app/services/fetch_link_card_service.rb
@@ -34,6 +34,14 @@ class FetchLinkCardService < BaseService
     nil
   end
 
+  ##
+  # Borrow most of this machinery to detect whether the status has at least one link.
+  def link?(status)
+    @status       = status
+    @original_url = parse_urls
+    !@original_url.nil?
+  end
+
   private
 
   def process_url