fix: paginate word mistakes by per_page instead of a hardcoded 100 (#721)

* fix: allow navigation to the last verse (6236) from draft editors

next_ayah_translation and next_ayah_tafsir both guard with < 6235, so when
the current record is verse 6235 (the second-to-last verse) the lookup is
skipped and editors cannot advance to the final verse 6236. The sibling
proofreading view (word_text_proofreadings/show.html.erb) correctly uses
< 6236. Fix both guards to use the true total of 6236 verses.

* fix: paginate word mistakes by per_page instead of a hardcoded 100

similar_words_data computed the offset and total_pages with a hardcoded
page size of 100, but the query itself uses .limit(per_page) where
per_page is 10 by default (max 50). With the default page size, page 2
already skipped past rows 11-100, and total_pages was wrong. Compute both
from per_page so pagination matches the actual page size.

Author: sporteka2

app/models/draft/tafsir.rb

@@ -292,7 +292,7 @@ def split_ayah_grouping
   end
 
   def next_ayah_tafsir
-    if end_verse_id < 6235
+    if end_verse_id < 6236
       Draft::Tafsir
         .where(resource_content_id: resource_content_id)
         .where("start_verse_id > ?", end_verse_id)

app/models/draft/translation.rb

@@ -59,7 +59,7 @@ def draft_text=(val)
   end
 
   def next_ayah_translation
-    if verse_id < 6235
+    if verse_id < 6236
       Draft::Translation
         .where(resource_content_id: resource_content_id)
         .where("verse_id > ?", verse_id)

app/presenters/word_mistakes_presenter.rb

@@ -91,7 +91,7 @@ def similar_words_data
                               .order('word_index ASC')
 
     page = [current_page, 1].max
-    offset = (page - 1) * 100
+    offset = (page - 1) * per_page
 
     similar_words = similar_words_query.offset(offset).limit(per_page)
     similar_mistakes = WordMistake
@@ -102,7 +102,7 @@ def similar_words_data
                          )
                          .group_by(&:word_id)
 
-    total_pages = (similar_words_query.count.to_f / 100).ceil
+    total_pages = (similar_words_query.count.to_f / per_page).ceil
 
     {
       count: similar_words_query.count,