Fixes incorrect result when the string includes unicode marks

minijinja-cli/Cargo.toml

@@ -46,7 +46,7 @@ minijinja = { version = "=2.5.0", path = "../minijinja", features = [
     "custom_syntax",
     "loop_controls"
 ] }
-minijinja-contrib = { version = "=2.5.0", optional = true, path = "../minijinja-contrib", features = ["pycompat", "datetime", "timezone", "rand", "unicode_wordwrap"] }
+minijinja-contrib = { version = "=2.5.0", optional = true, path = "../minijinja-contrib", features = ["pycompat", "datetime", "timezone", "rand", "unicode_wordwrap", "wordcount"] }
 rustyline = { version = "14.0.0", optional = true }
 serde = { version = "1.0.183", features = ["derive", "rc"] }
 serde_json = "1.0.105"

minijinja-contrib/Cargo.toml

@@ -21,6 +21,7 @@ pycompat = ["minijinja/builtins"]
 datetime = ["time"]
 timezone = ["time-tz"]
 rand = ["dep:rand"]
+wordcount = ["unicode_categories"]
 wordwrap = ["textwrap"]
 unicode_wordwrap = ["wordwrap", "textwrap/unicode-linebreak", "textwrap/unicode-width"]
 
@@ -31,6 +32,7 @@ serde = "1.0.164"
 textwrap = { version = "0.16.1", optional = true, default-features = false, features = ["smawk"] }
 time = { version = "0.3.35", optional = true, features = ["serde", "formatting", "parsing"] }
 time-tz = { version = "1.0.3", features = ["db"], optional = true }
+unicode_categories = { version = "0.1.1", optional = true}
 
 [dev-dependencies]
 insta = { version = "1.38.0", features = ["glob", "serde"] }

minijinja-contrib/src/filters/mod.rs

@@ -220,14 +220,18 @@ pub fn truncate(state: &State, value: Value, kwargs: Kwargs) -> Result<String, E
 /// ```jinja
 /// {{ "Hello world!"|wordcount }}
 /// ```
+#[cfg(feature = "wordcount")]
+#[cfg_attr(docsrs, doc(cfg(feature = "wordcount")))]
 pub fn wordcount(value: Value) -> Result<Value, Error> {
+    use unicode_categories::UnicodeCategories;
+
     let s = value.as_str().unwrap_or_default();
     let mut count: u32 = 0;
     let mut in_word = false;
 
     // Iterate through characters, counting transitions from non-word to word chars
     for c in s.chars() {
-        let is_word_char = c.is_alphanumeric() || c == '_';
+        let is_word_char = c.is_letter() || c.is_numeric() || c == '_';
         if is_word_char && !in_word {
             count += 1;
             in_word = true;

minijinja-contrib/src/lib.rs

@@ -35,7 +35,10 @@ pub fn add_to_environment(env: &mut Environment) {
     env.add_filter("pluralize", filters::pluralize);
     env.add_filter("filesizeformat", filters::filesizeformat);
     env.add_filter("truncate", filters::truncate);
-    env.add_filter("wordcount", filters::wordcount);
+    #[cfg(feature = "wordcount")]
+    {
+        env.add_filter("wordcount", filters::wordcount);
+    }
     #[cfg(feature = "wordwrap")]
     {
         env.add_filter("wordwrap", filters::wordwrap);

minijinja-contrib/tests/filters.rs

@@ -266,6 +266,18 @@ fn test_wordcount() {
         .unwrap(),
         "3"
     );
+
+    // Test unicode marks
+    assert_eq!(
+        env.render_str(
+            "{{ text|wordcount }}",
+            context! {
+                text => "helloाworld"
+            }
+        )
+        .unwrap(),
+        "2"
+    );
 }
 
 #[test]