Cv2 4582 no urls for similarity (#1885)

* Add test (should fail)

* Now reject strings that are only URLs too

---------

Co-authored-by: computermacgyver <computermacgyver>

app/models/bot/alegre.rb

@@ -1,4 +1,5 @@
 require 'open-uri'
+require 'uri'
 
 class Bot::Alegre < BotUser
   check_settings
@@ -20,7 +21,7 @@ class Error < ::StandardError
 
   REPORT_TEXT_SIMILARITY_FIELDS = ['report_text_title', 'report_text_content', 'report_visual_card_title', 'report_visual_card_content']
   ALL_TEXT_SIMILARITY_FIELDS = REPORT_TEXT_SIMILARITY_FIELDS + ['original_title', 'original_description', 'extracted_text', 'transcription', 'claim_description_content', 'fact_check_title', 'fact_check_summary']
-  BAD_TITLE_REGEX = /^[a-z\-]+-[0-9\-]+$/
+  BAD_TITLE_REGEX = /^[a-z\-]+-[0-9\-]+$|^#{URI.regexp}$/
   ::ProjectMedia.class_eval do
     attr_accessor :alegre_similarity_thresholds, :alegre_matched_fields
 

test/models/bot/alegre_4_test.rb

@@ -63,6 +63,16 @@ def teardown
     Bot::Alegre.unstub.stubs(:request)
   end
 
+  test "/text/similarity/ request should not be called for URL title" do
+    text = "http://meedan.com/@fun/thing?abc=def&hij=123"
+    pm1 = create_project_media team: @team, quote: text
+    Bot::Alegre.stubs(:request).raises("Request method called when it should not be")
+    assert_nothing_raised do
+        Bot::Alegre.send_to_text_similarity_index(pm1, create_field_instance, text, Bot::Alegre.item_doc_id(pm1, create_field_instance))
+    end
+    Bot::Alegre.unstub.stubs(:request)
+  end
+
   test "/text/similarity/ request should not be called for blank title" do
     text = ""
     pm1 = create_project_media team: @team, quote: text