fix(media): preserve Unicode prefixes in image-sequence filenames

Compute the final ASCII frame-number boundary without slicing inside a UTF-8 character. Preserve Unicode prefixes and add parser and public image-sequence import regressions, including missing frames and timeline placement.

Fixes #1565

Co-authored-by: Codex <[email protected]>

crates/engine/src/image_sequence_tests.rs

@@ -74,11 +74,20 @@ fn item_of(r: &Value) -> ItemId {
 
 #[test]
 fn import_as_image_sequence_with_missing_frames() {
-    let dir = tmp_dir("imgseq-import");
+    check_image_sequence_with_missing_frames("shot_");
+}
+
+#[test]
+fn import_unicode_image_sequence_with_missing_frames() {
+    check_image_sequence_with_missing_frames("é");
+}
+
+fn check_image_sequence_with_missing_frames(prefix: &str) {
+    let dir = tmp_dir(&format!("imgseq-import-{prefix}"));
     let mut s = demo();
     s.execute("prefs.set", json!({"key": "media.indeterminateTimebase", "value": "25"})).unwrap();
     // frames 1..=6 with 4 missing, plus an unrelated file
-    let paths = write_frames(&dir, "shot_", &[1, 2, 3, 5, 6], "png");
+    let paths = write_frames(&dir, prefix, &[1, 2, 3, 5, 6], "png");
     write_frames(&dir, "other_", &[1], "png");
     let r = s.execute("file.import", json!({"paths": [paths[0]], "imageSequence": true})).unwrap();
     let id = item_of(&r);

crates/media/README.md

@@ -0,0 +1,14 @@
+# filmcraft-media
+
+Media sources, still-image and image-sequence import, generated media, audio streams, and decoded
+frame caching for FilmCraft.
+
+## Image-sequence filenames
+
+`sequence::Numbered` identifies the last run of 1–18 ASCII frame-number digits in a filename's
+stem when the extension is a recognized still-image format. `/` and `\` separate directory
+components. A prefix may contain other UTF-8 characters, including accented letters, CJK and emoji:
+`é0001.png`, `图0001.png` and `🎞️0001.png`. Parsing preserves the prefix, number padding and suffix
+when generating sibling-frame paths. Sibling recognition compares the prefix's UTF-8 bytes exactly.
+Stems without ASCII digits or with a final frame-number run longer than 18 characters are not numbered
+image-sequence filenames.

crates/media/src/sequence.rs

@@ -46,7 +46,7 @@ impl Numbered {
         }
         let stem = &name[..ext_at];
         let end = stem.rfind(|c: char| c.is_ascii_digit())? + 1;
-        let start = stem[..end].rfind(|c: char| !c.is_ascii_digit()).map_or(0, |i| i + 1);
+        let start = stem[..end].trim_end_matches(|c: char| c.is_ascii_digit()).len();
         let digits = &stem[start..end];
         if digits.len() > 18 {
             return None;
@@ -271,6 +271,24 @@ mod tests {
         assert_eq!((w.dir.as_str(), w.prefix.as_str(), w.number), ("C:\\renders\\", "out.", 10));
     }
 
+    #[test]
+    fn numbered_unicode_prefix() {
+        for prefix in ["é", "图", "🎞️", "é1a_", "1é", "e\u{301}"] {
+            let path = format!("/shots/{prefix}0001.png");
+            let numbered = Numbered::parse(&path).unwrap();
+            assert_eq!(numbered.prefix, prefix);
+            assert_eq!(numbered.number, 1);
+            assert_eq!(numbered.digits, 4);
+            assert_eq!(numbered.path(2), format!("/shots/{prefix}0002.png"));
+            assert_eq!(numbered.number_of(&format!("{prefix}0002.png")), Some(2));
+        }
+        let numbered = Numbered::parse("/shots/é0001.png").unwrap();
+        assert_eq!(numbered.number_of("e\u{301}0002.png"), None, "prefixes must match exactly");
+        for name in ["é١٠٠١.png", "图０００１.png"] {
+            assert_eq!(Numbered::parse(name), None, "frame numbers require ASCII digits");
+        }
+    }
+
     #[test]
     fn frames_with_gaps() {
         let n = Numbered::parse("/d/s_003.png").unwrap();