From 79ba61b980b9fda015518e23df7528f47f1180f0 Mon Sep 17 00:00:00 2001
From: Maarten van Gompel <proycon@anaproy.nl>
Date: Tue, 10 May 2022 11:21:41 +0200
Subject: [PATCH] fix in outputting unicode offset #15 (for real now I hope)

---
 src/search.rs | 8 ++++----
 tests/main.rs | 4 ++--
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/src/search.rs b/src/search.rs
index 6f6a2e8..370ce7f 100644
--- a/src/search.rs
+++ b/src/search.rs
@@ -482,12 +482,12 @@ impl ContextRule {
 pub(crate) fn remap_offsets_to_unicodepoints<'a>(text: &'a str, mut matches: Vec<Match<'a>>) -> Vec<Match<'a>> {
     let mut bytes2unicodepoints: Vec<Option<usize>> = Vec::new();
     let mut end = 0;
-    for (unicodeoffset, (byteoffset, _char)) in text.char_indices().enumerate() {
-        for _ in bytes2unicodepoints.len()..byteoffset {
+    for (unicodeoffset, c) in text.chars().enumerate() {
+        bytes2unicodepoints.push(Some(unicodeoffset));
+        for _ in 0..c.len_utf8()-1 {
             bytes2unicodepoints.push(None);
         }
-        bytes2unicodepoints.push(Some(unicodeoffset));
-        end = byteoffset+1;
+        end = unicodeoffset+1;
     }
     //add an end offset
     bytes2unicodepoints.push(Some(end));
diff --git a/tests/main.rs b/tests/main.rs
index 4548d43..9551fe2 100644
--- a/tests/main.rs
+++ b/tests/main.rs
@@ -989,7 +989,7 @@ fn test0706_find_all_matches_unicodeoffsets() {
         model.add_to_vocabulary(text,None,&VocabParams::default());
     }
     model.build();
-    let matches = model.find_all_matches("I thиnk you are rihgt", &get_test_searchparams().with_max_ngram(1).with_unicodeoffsets());
+    let matches = model.find_all_matches("I thиnk you are righт", &get_test_searchparams().with_max_ngram(1).with_unicodeoffsets());
     assert!( !matches.is_empty() );
     assert_eq!( matches.get(0).unwrap().text , "I" );
     assert_eq!( matches.get(1).unwrap().text , "thиnk" );
@@ -998,7 +998,7 @@ fn test0706_find_all_matches_unicodeoffsets() {
     assert_eq!( model.match_to_str(matches.get(1).unwrap()) , "think" );
     assert_eq!( matches.get(2).unwrap().text , "you" );
     assert_eq!( matches.get(3).unwrap().text , "are" );
-    assert_eq!( matches.get(4).unwrap().text , "rihgt" );
+    assert_eq!( matches.get(4).unwrap().text , "righт" );
     assert_eq!( model.match_to_str(matches.get(4).unwrap()) , "right" );
 }