Fix wrapped-course false positive in picker; add build version display

Sort Unsorted's destination picker still listed a course whose real
video content sat two levels deep inside its own "display name"
folder (a common release-bundle extraction shape), alongside empty
leftover folders from the extraction that neither of the two earlier
exclusion checks could safely rule out - an empty folder is
indistinguishable from a legitimate freshly-created category. Detect
this case by comparing token similarity between a folder's own name
and its single course child's name: high overlap means they're naming
the same thing (a redundant wrapper), not a deliberate category
holding one course.

Also bake the build's git commit + build time into the Docker image
(git rev-parse at build time, .git deleted again immediately after) so
a push's effect on the running container is actually visible - shown
on the Settings page and in /health - instead of having to guess
whether a webhook rebuild picked up the latest commit.

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

.dockerignore

@@ -1,5 +1,8 @@
 # Version control
-.git
+# .git is intentionally NOT excluded - the Dockerfile reads it at build
+# time (git rev-parse) to bake the running commit into the image as
+# /app/BUILD_VERSION, then deletes .git itself in that same build step so
+# it never ends up in the final image.
 .gitignore
 .github
 

Dockerfile

@@ -5,8 +5,9 @@ FROM python:3.13.5-slim-bookworm
 WORKDIR /app
 
 # ffprobe (from ffmpeg) reads each video's duration for the library
-# browser's "how long is this course" display
+# browser's "how long is this course" display; git is only needed
-RUN apt-get update && apt-get install -y --no-install-recommends ffmpeg \
+# transiently below, to bake this build's commit into the image
+RUN apt-get update && apt-get install -y --no-install-recommends ffmpeg git \
     && rm -rf /var/lib/apt/lists/*
 
 # copy the dependencies file to the working directory
@@ -19,6 +20,15 @@ RUN pip install -r requirements.txt
 # copy the content of the local src directory to the working directory
 COPY . .
 
+# Record which commit and when this image was built, so the running app
+# can show it (Settings page, /health) - a full-image-rebuild webhook
+# deploy gives no other visible confirmation that a push actually landed
+# in the running container. .git is removed right after so the image
+# doesn't carry the repo's full history.
+RUN (git rev-parse --short HEAD 2>/dev/null || echo unknown) > /app/BUILD_VERSION \
+    && date -u +"%Y-%m-%d %H:%M UTC" >> /app/BUILD_VERSION \
+    && rm -rf /app/.git
+
 EXPOSE 5000
 
 # add healthcheck using Python standard library

README.md

@@ -44,6 +44,14 @@ triggers a full image **rebuild** from the Dockerfile (confirmed, not just a
 container restart), so Dockerfile changes (e.g. adding `ffmpeg`) take effect
 on the very next push without any manual step.
 
+**Confirming a deploy landed:** the Dockerfile bakes the exact commit and
+build time into the image as `/app/BUILD_VERSION` (`git rev-parse` at build
+time, `.git` itself is deleted again right after so it doesn't ship in the
+image) - shown at the bottom of the Settings page and in `/health`'s JSON
+response, so after a push you can check the running container actually
+picked it up instead of guessing. Reads as `dev` outside Docker (no
+`BUILD_VERSION` file to read).
+
 ---
 
 ## Local development
@@ -109,12 +117,15 @@ silently stay blank instead of erroring.
   the same picker Manage Library's Move action uses. The picker excludes
   actual course/item folders, not just organizational ones: a leaf folder
   holding a single non-video/audio file (an ebook, an audiobook in a
-  format this app doesn't play, etc.), and a folder whose whole subtree
+  format this app doesn't play, etc.); a folder whose whole subtree has no
-  has no video/audio anywhere in it at all (a course's bundled source
+  video/audio anywhere in it at all (a course's bundled source code, a
-  code, a Python virtualenv, project assets, ...) even though it has
+  Python virtualenv, project assets, ...) even though it has plenty of
-  plenty of subfolders - both cases Library browsing's own course
+  subfolders; and a release-bundle folder that wraps a single real course
-  detection wouldn't catch either, since there's no video/audio file to
+  one level down under its own "course display name" folder, alongside
-  key off. Nothing on disk moves until you review and hit Apply. Matching
+  unrelated (possibly empty) junk siblings at the same level - all three
+  cases Library browsing's own course detection wouldn't catch either,
+  since it only checks one level of nesting for video/audio. Nothing on
+  disk moves until you review and hit Apply. Matching
   ignores common noise (e-learning
   platform names, release/distribution-group tags, dates) via a stopword
   list, and beyond that treats a match against a folder's own deliberate

offlineu_core.py

@@ -30,6 +30,32 @@ from flask import Flask, render_template, request, jsonify, send_file, redirect,
 app = Flask(__name__)
 app.config['SECRET_KEY'] = 'your-secret-key-change-in-production'
 
+
+def _load_build_version() -> str:
+    """
+    Commit + build time baked into the image at build time (see
+    Dockerfile) - lets Settings and /health show which version is
+    actually running, the only way to confirm a git push made it into a
+    rebuilt container, since the webhook does a full image rebuild with
+    no other visible confirmation. Falls back to 'dev' outside Docker,
+    where there's no BUILD_VERSION file (e.g. running via --library-path
+    directly).
+    """
+    version_file = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'BUILD_VERSION')
+    try:
+        with open(version_file, 'r') as f:
+            lines = [line.strip() for line in f if line.strip()]
+        if len(lines) >= 2:
+            return f'{lines[0]} · built {lines[1]}'
+        if lines:
+            return lines[0]
+    except OSError:
+        pass
+    return 'dev'
+
+
+BUILD_VERSION = _load_build_version()
+
 # Supported file types
 VIDEO_EXTENSIONS = {'.mp4', '.mkv', '.avi', '.mov', '.webm', '.m4v', '.flv', '.wmv'}
 AUDIO_EXTENSIONS = {'.mp3', '.wav', '.m4a', '.aac', '.ogg', '.flac'}
@@ -1324,6 +1350,41 @@ def _is_media_free_subtree(directory: Path) -> bool:
     return not _subtree_has_any_media(directory)
 
 
+def _is_course_wrapper(directory: Path, min_similarity: float = 0.5) -> bool:
+    """
+    A release-bundle folder whose name is really just a course title with
+    extra decoration (platform name, distribution tag, ...), wrapping a
+    single actual course folder one level down alongside unrelated junk
+    at the same level - e.g. a "Udemy...BOOKWARE-XYZ" folder containing
+    both "Course Title Here/Chapter 1/video.mp4" *and* a pile of bundled
+    source code/empty leftover folders from extraction as siblings.
+    _looks_like_course only checks one level of nesting, so this reads as
+    a bare category with a course buried inside it - and unlike
+    _is_media_free_subtree, the junk siblings here can be perfectly
+    ordinary empty folders (indistinguishable from an intentionally empty,
+    freshly-created category), so that check alone doesn't catch it.
+    Detected by token similarity (the same Jaccard measure
+    _find_duplicate_courses uses) between the directory's own name and its
+    one course child's name: high overlap means they're naming the same
+    thing, not a deliberate parent/child relationship like "IT/AWS"
+    holding an "AWS Certified..." course, whose names barely overlap.
+    """
+    try:
+        children = [c for c in directory.iterdir() if c.is_dir() and not c.name.startswith('.')]
+    except (PermissionError, OSError):
+        return False
+    course_children = [c for c in children if _looks_like_course(c)]
+    if len(course_children) != 1:
+        return False
+    dir_tokens = _tokenize(directory.name)
+    course_tokens = _tokenize(course_children[0].name)
+    union_tokens = dir_tokens | course_tokens
+    if not union_tokens:
+        return False
+    similarity = len(dir_tokens & course_tokens) / len(union_tokens)
+    return similarity >= min_similarity
+
+
 def _build_category_index() -> List[Dict[str, Any]]:
     """
     Every category/subcategory folder currently in the library (e.g. IT,
@@ -1345,12 +1406,16 @@ def _build_category_index() -> List[Dict[str, Any]]:
     Stops recursing into a folder once it reads as a course itself
     (_looks_like_course, the same rule the Library browser uses), an
     unrecognized leaf item (_is_unrecognized_leaf_item - an ebook/
-    audiobook/misc file that isn't video or audio), or a subtree with no
+    audiobook/misc file that isn't video or audio), a subtree with no
     video/audio anywhere in it at all (_is_media_free_subtree - a course's
-    bundled source code/project files, a Python virtualenv, etc.), so
+    bundled source code/project files, a Python virtualenv, etc.), or a
-    individual courses and their non-lecture contents never show up as if
+    release-bundle wrapper around a single course one level down
-    they were categories to file things under. The Unsorted folder itself
+    (_is_course_wrapper - the video content sits under an extra "course
-    is excluded - it's the source, never a valid destination.
+    display name" folder, one level deeper than _looks_like_course checks,
+    alongside unrelated - possibly empty - junk siblings), so individual
+    courses and their non-lecture contents never show up as if they were
+    categories to file things under. The Unsorted folder itself is
+    excluded - it's the source, never a valid destination.
     """
     library_root = Path(get_library_root())
     try:
@@ -1373,7 +1438,8 @@ def _build_category_index() -> List[Dict[str, Any]]:
                     continue
             except OSError:
                 pass
-            if _looks_like_course(entry) or _is_unrecognized_leaf_item(entry) or _is_media_free_subtree(entry):
+            if (_looks_like_course(entry) or _is_unrecognized_leaf_item(entry)
+                    or _is_media_free_subtree(entry) or _is_course_wrapper(entry)):
                 continue
             new_parts = path_parts + [entry.name]
             path_tokens: Set[str] = set()
@@ -3633,6 +3699,7 @@ def settings_page():
         card_style_choices=['flat', 'elevated', 'bordered'],
         corner_radius_choices=['sharp', 'rounded', 'pill'],
         active_tab='settings',
+        build_version=BUILD_VERSION,
     )
 
 
@@ -4365,7 +4432,7 @@ def serve_file(filepath):
 @app.route('/health')
 def healthcheck():
     """Healthcheck endpoint for Docker"""
-    return jsonify({"status": "healthy"}), 200
+    return jsonify({"status": "healthy", "version": BUILD_VERSION}), 200
 
 @app.route('/reset_course')
 def reset_course():

templates/settings.html

@@ -466,6 +466,10 @@
             <button class="btn btn-secondary" onclick="resetSettings()">Reset to Defaults</button>
             <span id="save-status">{{ icons.icon('check', 14) }} Saved</span>
         </div>
+
+        <p style="text-align: center; color: var(--text-muted); font-size: 0.8em; margin-top: 30px;">
+            Version {{ build_version }}
+        </p>
     </div>
     </div>