Image sets: suffixes that don't count the images go first

The set bound cut at an image name's last number, but in many threads
that's a suffix, so every image was a set ("It does my heart good to see
Red Dog dancing again": 4,683 releases of 4,730 files): a repeat of the
index (katelg010_10), a year (hgrscans1097_2009), a size
(pj02kristina_24of40_1600), an "of" total (24of40). imageSet drops
those, then cuts at the last number (cutie-071-032 stays cutie-071-).
Both implementations; tests; fixtures identical.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

edge/src/names.rs

@@ -414,18 +414,48 @@ static AUDIO: LazyLock<Regex> = LazyLock::new(|| re(r"(?i)\.(mp3|flac|m4a|m4b|og
 static DOCUMENT: LazyLock<Regex> = LazyLock::new(|| re(r"(?i)\.(pdf|epub|mobi|azw3?|djvu|cbr|cbz|cb7|chm|lit|docx?)$"));
 
 /// An image set's name without a counter: up to the last number, without numbers.
-/// `/\d/.test(stem) ? stem.replace(/\d+\D*$/, '') : ''`: an image's set, digits kept.
-fn image_set(stem: &str) -> String {
-    let b = stem.as_bytes();
-    match b.iter().rposition(u8::is_ascii_digit) {
+/// imageSet (lib/release/identity.ts): the name up to the number that counts the images, digits kept.
+pub fn image_set(stem: &str) -> String {
+    // .replace(/(\d+)\s*of\s*\d+/gi, '$1')
+    let mut s = OF_TOTAL.replace_all(stem.as_bytes(), &b"$1"[..]).into_owned();
+    loop {
+        // /[_-](\d+)$/
+        let digits = s.iter().rev().take_while(|c| c.is_ascii_digit()).count();
+        if digits == 0 || s.len() <= digits || !matches!(s[s.len() - digits - 1], b'_' | b'-') {
+            break;
+        }
+        let n = &s[s.len() - digits..];
+        let before = &s[..s.len() - digits - 1];
+        // /(\d+)\D*$/: the last digit run before
+        let Some(last) = before.iter().rposition(u8::is_ascii_digit) else { break };
+        let start = before[..=last].iter().rposition(|c| !c.is_ascii_digit()).map_or(0, |i| i + 1);
+        let previous = &before[start..=last];
+        let same = trim_zeros(n) == trim_zeros(previous);
+        let year = n.len() == 4 && (n.starts_with(b"19") || n.starts_with(b"20"));
+        if same || year || n.len() >= 4 {
+            let cut = before.len();
+            s.truncate(cut);
+        } else {
+            break;
+        }
+    }
+    match s.iter().rposition(u8::is_ascii_digit) {
         Some(last) => {
-            let start = b[..=last].iter().rposition(|c| !c.is_ascii_digit()).map_or(0, |i| i + 1);
-            stem[..start].to_string()
+            let start = s[..=last].iter().rposition(|c| !c.is_ascii_digit()).map_or(0, |i| i + 1);
+            String::from_utf8_lossy(&s[..start]).into_owned()
         }
         None => String::new(),
     }
 }
 
+/// Number(digits) equality for digit strings: without leading zeros (all zeros: "0").
+fn trim_zeros(d: &[u8]) -> &[u8] {
+    let t = d.iter().position(|&c| c != b'0').map_or(&d[d.len().saturating_sub(1)..], |i| &d[i..]);
+    t
+}
+
+static OF_TOTAL: LazyLock<Regex> = LazyLock::new(|| re(r"(?i)([0-9]+)SP*ofSP*[0-9]+"));
+
 fn image_prefix(stem: &str) -> String {
     let b = stem.as_bytes();
     let cut = match b.iter().rposition(u8::is_ascii_digit) {
@@ -1130,6 +1160,25 @@ mod image_set_tests {
         assert!(!random_poster("w.merklein@freenet.de (muffin)"));
     }
 
+    #[test]
+    fn image_sets_without_suffixes() {
+        use super::image_set;
+        for (stem, want) in [
+            ("stp250x001", "stp250x"),
+            ("cutie-071-032", "cutie-071-"),
+            ("katelg010_10", "katelg"),
+            ("kate112lg035_35", "kate112lg"),
+            ("!!hgrscans1097_2009", "!!hgrscans"),
+            ("pj02kristina_24of40_1600", "pj02kristina_"),
+            ("keira_s04_l keira_s04_0040_l", "keira_s04_l keira_s04_"),
+            ("IMG_2676", "IMG_"),
+            ("2877-041", "2877-"),
+            ("JimAndJoey Kindly Husband-to-Be", ""),
+        ] {
+            assert_eq!(image_set(stem), want, "{stem}");
+        }
+    }
+
     #[test]
     fn image_sets_by_their_numbered_name() {
         let silver = |n: &str| release_identity("p", &format!(r#"The excellent STIRILING SILVER - Always popular - "{n}" yEnc 320794 bytes"#), n);

lib/release/identity.ts

@@ -141,6 +141,27 @@ const DOCUMENT = /\.(pdf|epub|mobi|azw3?|djvu|cbr|cbz|cb7|chm|lit|docx?)$/i
 /** Dates in names: 2011-08-29, 08-29-2011 or 29.08.2011, 02-2011. */
 const DATES = /\b(?:(?:19|20)\d{2}[-._ ](?:0?[1-9]|1[0-2])[-._ ]\d{1,2}|\d{1,2}[-._ ]\d{1,2}[-._ ](?:19|20)\d{2}|\d{1,2}[-._](?:19|20)\d{2})\b/g
 
+/**
+ * An image's set: its name up to the number that counts the images, digits before it kept ("stp250x001" →
+ * "stp250x", "cutie-071-032" → "cutie-071-"); empty for a name without numbers. Suffixes after it that don't
+ * count the images go first: a repeat of the number before (katelg010_10), a year (hgrscans1097_2009), a
+ * size of 4+ digits (…_24of40_1600), and an "of" total (24of40) (LOG.md 2026-10-12 (3)).
+ */
+export function imageSet(stem: string): string {
+  let s = stem.replace(/(\d+)\s*of\s*\d+/gi, '$1')
+  while (true) {
+    const m = /[_-](\d+)$/.exec(s)
+    if (!m) break
+    const before = s.slice(0, m.index)
+    const previous = /(\d+)\D*$/.exec(before)
+    if (!previous) break
+    const n = m[1]
+    if (Number(n) === Number(previous[1]) || /^(?:19|20)\d\d$/.test(n) || n.length >= 4) s = before
+    else break
+  }
+  return /\d/.test(s) ? s.replace(/\d+\D*$/, '') : ''
+}
+
 export function counterTotal(subject: string): number | undefined {
   const s = subject.replace(/"[^"]*"/g, ' ')
   const plain = /\[(\d+)\/(\d+)\]|\((\d+)\/(\d+)\)/.exec(s)
@@ -257,7 +278,7 @@ export function releaseIdentity(input: ReleaseInput): ReleaseIdentity | null {
     // Without a counter an image's set bounds it: its name up to its last number, digits kept ("stp250x001.jpg"
     // → "stp250x": STIRILING SILVER's 300 sets were one release of 44k images); a name without numbers names no
     // set ("Kindly Husband-to-Be.JPG": a thread of 1,177 releases), so the subject alone (LOG.md 2026-10-12 (2)).
-    const set = image && /\d/.test(stem) ? stem.replace(/\d+\D*$/, '') : ''
+    const set = image ? imageSet(stem) : ''
     return {
       key: hash('subject', poster, remainder, total ?? set),
       name: remainder,

tests/unit/release/identity.spec.ts

@@ -222,6 +222,22 @@ test.group('release / par2 names', () => {
     assert.equal(mint('JimAndJoey Kindly Husband-to-Be.JPG').key, mint('Oops OnlyOneMore.JPG').key)
   })
 
+  test("an image's set: suffixes that don't count the images go first", async ({ assert }) => {
+    const { imageSet } = await import('#lib/release/identity')
+    for (const [stem, set] of [
+      ['stp250x001', 'stp250x'],
+      ['cutie-071-032', 'cutie-071-'],
+      ['katelg010_10', 'katelg'],
+      ['kate112lg035_35', 'kate112lg'],
+      ['!!hgrscans1097_2009', '!!hgrscans'],
+      ['pj02kristina_24of40_1600', 'pj02kristina_'],
+      ['keira_s04_l keira_s04_0040_l', 'keira_s04_l keira_s04_'],
+      ['IMG_2676', 'IMG_'],
+      ['2877-041', '2877-'],
+      ['JimAndJoey Kindly Husband-to-Be', ''],
+    ]) assert.equal(imageSet(stem), set, stem)
+  })
+
   test('dates are no file counters', async ({ assert }) => {
     const { counterTotal } = await import('#lib/release/identity')
     assert.isUndefined(counterTotal('Avira AntiVir Rescue System 08-29-2011.rar'))