From e6cfa7ab897ac4fd6bf9bbcafe09fc42c4cc2c5d Mon Sep 17 00:00:00 2001
From: Eugen Rochko <eugen@zeonfederated.com>
Date: Tue, 26 Mar 2019 01:23:59 +0100
Subject: Change language detector threshold from 140 characters to 4 words
 (#10376)

Add `lang` attribute to statuses in web UI
---
 app/lib/language_detector.rb | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'app/lib')

diff --git a/app/lib/language_detector.rb b/app/lib/language_detector.rb
index 70a9084d1..1e90af42d 100644
--- a/app/lib/language_detector.rb
+++ b/app/lib/language_detector.rb
@@ -3,7 +3,7 @@
 class LanguageDetector
   include Singleton
 
-  CHARACTER_THRESHOLD    = 140
+  WORDS_THRESHOLD        = 4
   RELIABLE_CHARACTERS_RE = /[\p{Hebrew}\p{Arabic}\p{Syriac}\p{Thaana}\p{Nko}\p{Han}\p{Katakana}\p{Hiragana}\p{Hangul}]+/m
 
   def initialize
@@ -37,7 +37,7 @@ class LanguageDetector
   end
 
   def sufficient_text_length?(text)
-    text.size >= CHARACTER_THRESHOLD
+    text.split(/\s+/).size >= WORDS_THRESHOLD
   end
 
   def language_specific_character_set?(text)
-- 
cgit