Fix course-detection heuristic swallowing category folders as one course

_looks_like_course() treated any folder with 2+ media-holding subfolders
as a single multi-section course, with no check that those subfolders
actually looked like sections. A category folder holding many unrelated
courses (e.g. Claude/, General/) could accidentally satisfy that if
several of its courses happened to keep episodes directly at their own
top level rather than nested under a chapter folder, collapsing the
whole category into one bogus course.

Now every media-holding subfolder must read as a chapter/section label
(Section 1, Module 2, or a numbered chapter like 01-Introduction) for
the parent to count as one course - verified against the real library
mount that ChatGPT/Claude/Copilot/General all correctly show as
folders again, while individually-structured courses still detect
correctly.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

offlineu_core.py

@@ -526,25 +526,26 @@ def find_course_thumbnail(course_path: str) -> Optional[str]:
 
 
 _SECTION_NAME_RE = re.compile(
-    r'^(section|module|chapter|part|unit|lesson)\b', re.IGNORECASE
+    r'^(section|module|chapter|part|unit|lesson)\b|^\d+[\s_.\-]', re.IGNORECASE
 )
 
 
 def _looks_like_course(directory: Path) -> bool:
     """
-    Heuristic for 'this folder is a course, stop recursing into it':
+    Heuristic for 'this folder is a course, stop recursing into it': it has
-    it has media files directly, or at least two of its immediate
+    media files directly, or its media-holding immediate subfolders all
-    subfolders do (covers the common Section 1/, Section 2/... layout).
+    read as chapter/section labels ("Section 1", "Module 2", "01 -
+    Introduction", ...) rather than course titles - covers the common
+    Section 1/, Section 2/... (or numbered-chapter) layout.
 
-    Requiring 2+ matching subfolders (rather than just 1) avoids mistaking
+    Requiring *every* media-holding subfolder to match, not just one, is
-    a publisher/grouping folder that holds a single course - e.g.
+    what keeps a publisher/category folder holding several unrelated
-    Pluralsight/Docker Deep Dive/lesson.mp4 - for the course itself.
+    courses - e.g. Claude/Pluralsight.X.../*.mp4 next to
-
+    Claude/Linkedin.Learning.Y.../*.mp4 - from being mistaken for a single
-    The one exception: a directory with exactly one media-holding subfolder
+    course just because more than one of its subfolders happens to store
-    whose name reads as a section label ("Section 1", "Module 2", ...)
+    episodes directly rather than nested under their own chapter folder.
-    rather than a course title. That's still a course with just one
+    Real course titles ("Pluralsight.Docker.Deep.Dive...") don't read as
-    section, not a publisher folder - Pluralsight/Docker Deep Dive doesn't
+    chapter labels, so this doesn't reopen that ambiguity.
-    get named "Section 1", so this doesn't reopen the ambiguity above.
     """
     if _has_direct_media(directory):
         return True
@@ -556,14 +557,9 @@ def _looks_like_course(directory: Path) -> bool:
     except (PermissionError, OSError):
         return False
     children_with_media = [child for child in children if _has_direct_media(child)]
-    if len(children_with_media) >= 2:
+    if not children_with_media:
-        return True
+        return False
-    if (
+    return all(_SECTION_NAME_RE.match(child.name.strip()) for child in children_with_media)
-        len(children_with_media) == 1
-        and _SECTION_NAME_RE.match(children_with_media[0].name.strip())
-    ):
-        return True
-    return False
 
 
 HIDDEN_PATHS_FILE = os.path.join(DATA_DIR, 'hidden_paths.json')