diff --git a/CHANGELOG.md b/CHANGELOG.md
index 6cf63ee..03208f7 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,3 +1,4 @@
+2026-08-25 12:25 UTC — Remove unreliable duplicate-lesson-file matching
2026-08-25 00:33 UTC — Widen thumbnail sampling window past platform bumpers
2026-08-25 00:26 UTC — Fix race condition in thumbnail candidate generation
2026-08-25 00:20 UTC — Add autoplay, progress rings, duplicate-lesson detection, title-card thumbnails
diff --git a/README.md b/README.md
index a117652..142fcbc 100644
--- a/README.md
+++ b/README.md
@@ -185,20 +185,17 @@ silently stay blank instead of erroring.
difference doesn't hide a real duplicate. Groups by similarity (union of
any two courses over the threshold, transitively) rather than showing
raw pairs; each course in a group gets a one-click Hide, or a **Delete**
- that permanently removes it from disk (confirmed with the full path
- before it runs). A group can also be marked **"Not a duplicate"** if the
- match is wrong, which excludes that specific pair from future scans
- without touching anything else that happens to match one of those
- courses; ignored pairs are listed (and reversible) under "Ignored
- matches," and persist through backup/restore alongside hidden paths and
- Next Up.
- - *Duplicate Lesson Files*: the same idea one level down - media files
- sitting in the same folder that look like the same lesson downloaded
- twice (e.g. a re-download that landed alongside the original instead of
- replacing it). Compares only within a folder, at a higher match
- threshold than Duplicate Courses, so two different lessons on a similar
- topic don't get flagged; shares its ignore list with Duplicate Courses.
- Delete here removes a single file, not a whole course.
+ that permanently removes it from disk (the one destructive action in the
+ whole app - confirmed with the full path before it runs). A group can
+ also be marked **"Not a duplicate"** if the match is wrong, which
+ excludes that specific pair from future scans without touching anything
+ else that happens to match one of those courses; ignored pairs are
+ listed (and reversible) under "Ignored matches," and persist through
+ backup/restore alongside hidden paths and Next Up. (Matching is
+ deliberately scoped to whole course folders, not individual lesson
+ files - courses that split one topic across several numbered files,
+ e.g. "...Part 1"/"...Part 2", look just as similar by name as a genuine
+ re-download, so file-level matching wasn't reliable enough to keep.)
- *Clean Up Stale References*: finds entries in the hidden-paths list,
Next Up queue, Favorites, or Recently Viewed history that point at a path
no longer on disk - normally from renaming/moving/deleting a course
diff --git a/VERSION b/VERSION
index 17c78b1..8658d3e 100644
--- a/VERSION
+++ b/VERSION
@@ -1 +1 @@
-2026-08-25 00:33 UTC — widen thumbnail sampling window past platform bumpers
+2026-08-25 12:25 UTC — remove unreliable duplicate-lesson-file matching
diff --git a/offlineu_core.py b/offlineu_core.py
index e610e3f..c0bb4ae 100644
--- a/offlineu_core.py
+++ b/offlineu_core.py
@@ -1864,64 +1864,6 @@ def _find_duplicate_courses(min_similarity: float = 0.6) -> List[Dict[str, Any]]
return groups
-def _find_duplicate_lesson_files(min_similarity: float = 0.75) -> List[Dict[str, Any]]:
- """
- Within each course, media files sitting in the same folder whose
- names look like the same lesson downloaded twice - e.g. "03 Setting
- Up Your Environment.mp4" and "03 Setting Up Your Environment (1).mp4"
- left behind by a re-download that landed in the same section. Scoped
- to siblings in the same folder, unlike Duplicate Courses' whole-library
- comparison: two genuinely different lessons can share a lot of
- vocabulary ("Part 1"/"Part 2" of a topic) without being duplicates, so
- same-folder pairing plus a higher similarity threshold keeps this
- conservative. Shares Duplicate Courses' ignore list (get_ignored_
- duplicate_pairs) - a pair is just a pair of paths there, whether
- they're course folders or files. Read-only.
- """
- ignored_pairs = get_ignored_duplicate_pairs()
- groups = []
- for course_dir in get_all_course_dirs():
- by_folder: Dict[Path, List[Path]] = {}
- for f in course_dir.rglob('*'):
- if f.is_file() and f.suffix.lower() in VIDEO_EXTENSIONS | AUDIO_EXTENSIONS:
- by_folder.setdefault(f.parent, []).append(f)
-
- for files in by_folder.values():
- if len(files) < 2:
- continue
- entries = [{'path': f, 'tokens': _tokenize(f.stem)} for f in files]
- n = len(entries)
- for i in range(n):
- for j in range(i + 1, n):
- path_i, path_j = str(entries[i]['path']), str(entries[j]['path'])
- if tuple(sorted((path_i, path_j))) in ignored_pairs:
- continue
- a, b = entries[i]['tokens'], entries[j]['tokens']
- union_tokens = a | b
- if not union_tokens:
- continue
- similarity = len(a & b) / len(union_tokens)
- if similarity < min_similarity:
- continue
- try:
- size_i = entries[i]['path'].stat().st_size
- size_j = entries[j]['path'].stat().st_size
- except OSError:
- continue
- groups.append({
- 'course_name': course_dir.name,
- 'course_path': str(course_dir),
- 'files': sorted([
- {'path': path_i, 'name': entries[i]['path'].name, 'bytes': size_i, 'human': _format_bytes(size_i)},
- {'path': path_j, 'name': entries[j]['path'].name, 'bytes': size_j, 'human': _format_bytes(size_j)},
- ], key=lambda f: f['name'].lower()),
- 'similarity': round(similarity, 2),
- })
-
- groups.sort(key=lambda g: g['similarity'], reverse=True)
- return groups
-
-
def _find_stale_references() -> Dict[str, List[Dict[str, str]]]:
"""
Entries in hidden_paths.json / next_up.json / recent_views.json that
@@ -3760,47 +3702,6 @@ def undo_last_action_api():
})
-@app.route('/api/library/duplicate-lessons')
-def duplicate_lesson_files_api():
- """Scan for lesson files within the same course/folder that look like
- the same lesson downloaded twice (see _find_duplicate_lesson_files).
- Manual trigger, same reasoning as Duplicate Courses."""
- groups = _find_duplicate_lesson_files()
- return jsonify({'groups': groups})
-
-
-@app.route('/api/library/delete-file', methods=['POST'])
-def delete_library_file_api():
- """
- Permanently delete a single media file from disk - the file-level
- counterpart to /api/library/delete, used by Duplicate Lesson Files.
- Scoped to files (not directories) inside the library root with a
- recognized media extension, so it can't be pointed at something else.
- """
- data = request.json or {}
- path = (data.get('path') or '').strip()
- if not path:
- return jsonify({'success': False, 'error': 'Missing path'}), 400
-
- library_root = os.path.abspath(get_library_root())
- target_abs = os.path.abspath(path)
-
- if not target_abs.startswith(library_root + os.sep):
- return jsonify({'success': False, 'error': 'Path is outside the library'}), 403
- if not os.path.isfile(target_abs):
- return jsonify({'success': False, 'error': 'No longer exists - already deleted?'}), 404
- if Path(target_abs).suffix.lower() not in VIDEO_EXTENSIONS | AUDIO_EXTENSIONS:
- return jsonify({'success': False, 'error': 'Not a media file'}), 400
-
- try:
- os.remove(target_abs)
- except OSError as e:
- return jsonify({'success': False, 'error': str(e)}), 500
-
- invalidate_cache()
- return jsonify({'success': True})
-
-
@app.route('/api/library/duplicates')
def duplicate_courses_api():
"""Scan for courses that look like copies of each other by name (see
diff --git a/templates/unsorted.html b/templates/unsorted.html
index 1d63112..576179b 100644
--- a/templates/unsorted.html
+++ b/templates/unsorted.html
@@ -687,16 +687,6 @@
-
Scan for lesson files sitting in the same folder that look like the same lesson downloaded twice - a re-download that landed alongside the original instead of replacing it. Compares files within each folder only, with a higher match threshold than Duplicate Courses, so two genuinely different lessons on a similar topic aren't flagged.
-
-
-
-
-
-
-
-
{{ icons.icon('trash', 18) }} Clean Up Stale References
Renaming, moving, or deleting a course directly on the NAS (instead of through this app) can leave behind references in Hidden, Next Up, or Recently Viewed that point at nothing. Review and remove them here - course files on disk are never touched.