Integrate FilterEngine with app.py (Phase 4)

Complete integration of filter pipeline with web application:

App.py Integration:
- Initialize FilterEngine singleton at startup
- Update /api/posts endpoint to use FilterEngine.apply_filterset()
- Apply user's filterset preference from settings
- Sort posts by filter_score (highest first), then timestamp
- Add filter metadata to post responses (filter_score, categories, tags)

Settings Page Updates:
- Dynamically load available filtersets from FilterEngine
- Show filterset descriptions in settings UI
- Validate filterset selection against FilterEngine

Security:
- Update _is_safe_filterset() to use FilterEngine's list
- Dynamic ALLOWED_FILTERSETS from filtersets.json

User Experience:
- Posts automatically filtered based on user preferences
- Quality/relevance scores affect post ordering
- Transparent filter metadata available in API

Caching:
- FilterEngine uses 3-level cache for efficiency
- Cache reused across page loads (5min TTL)
- AI results cached permanently

Next Steps:
- Polling service integration
- Database model for persistent results
- UI for cache stats and filter debugging

Related to filtering engine implementation

🤖 Generated with [Claude Code](https://claude.com/claude-code)

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

app.py

@@ -51,7 +51,7 @@ app.config['AUTH0_CLIENT_SECRET'] = os.getenv('AUTH0_CLIENT_SECRET', '')
 app.config['AUTH0_AUDIENCE'] = os.getenv('AUTH0_AUDIENCE', '')
 
 # Configuration constants
-ALLOWED_FILTERSETS = {'no_filter', 'safe_content'}
+# Note: ALLOWED_FILTERSETS will be dynamically loaded from filter_engine
 ALLOWED_EXTENSIONS = {'png', 'jpg', 'jpeg', 'gif'}
 UPLOAD_FOLDER = 'static/avatars'
 MAX_FILENAME_LENGTH = 100
@@ -82,6 +82,11 @@ from polling_service import polling_service
 polling_service.init_app(app)
 polling_service.start()
 
+# Initialize filter engine
+from filter_pipeline import FilterEngine
+filter_engine = FilterEngine.get_instance()
+logger.info(f"FilterEngine initialized with {len(filter_engine.get_available_filtersets())} filtersets")
+
 # Initialize OAuth for Auth0
 oauth = OAuth(app)
 auth0 = oauth.register(
@@ -105,7 +110,9 @@ def _is_safe_filterset(filterset):
     """Validate filterset name for security"""
     if not filterset or not isinstance(filterset, str):
         return False
-    return filterset in ALLOWED_FILTERSETS and re.match(r'^[a-zA-Z0-9_-]+$', filterset)
+    # Check against available filtersets from filter_engine
+    allowed_filtersets = set(filter_engine.get_available_filtersets())
+    return filterset in allowed_filtersets and re.match(r'^[a-zA-Z0-9_-]+$', filterset)
 
 def _is_safe_path(path):
     """Validate file path for security"""
@@ -356,24 +363,31 @@ def api_posts():
         platform = request.args.get('platform', '')
         search_query = request.args.get('q', '').lower().strip()
 
+        # Get user's filterset preference (or default to no_filter)
+        filterset_name = 'no_filter'
+        if current_user.is_authenticated:
+            try:
+                user_settings = json.loads(current_user.settings) if current_user.settings else {}
+                filterset_name = user_settings.get('filter_set', 'no_filter')
+            except:
+                filterset_name = 'no_filter'
+
         # Use cached data for better performance
         cached_posts, cached_comments = _load_posts_cache()
 
-        posts = []
+        # Collect raw posts for filtering
-        
+        raw_posts = []
-        # Process cached posts
         for post_uuid, post_data in cached_posts.items():
-            # Apply community filter
+            # Apply community filter (before filterset)
             if community and post_data.get('source', '').lower() != community.lower():
                 continue
 
-            # Apply platform filter
+            # Apply platform filter (before filterset)
             if platform and post_data.get('platform', '').lower() != platform.lower():
                 continue
 
-            # Apply search filter
+            # Apply search filter (before filterset)
             if search_query:
-                # Search in title, content, author, and source
                 title = post_data.get('title', '').lower()
                 content = post_data.get('content', '').lower()
                 author = post_data.get('author', '').lower()
@@ -385,7 +399,15 @@ def api_posts():
                        search_query in source):
                     continue
 
-            # Get comment count from cache
+            raw_posts.append(post_data)
+
+        # Apply filterset using FilterEngine
+        filtered_posts = filter_engine.apply_filterset(raw_posts, filterset_name, use_cache=True)
+
+        # Build response posts with metadata
+        posts = []
+        for post_data in filtered_posts:
+            post_uuid = post_data.get('uuid')
             comment_count = len(cached_comments.get(post_uuid, []))
 
             # Get proper display name for source
@@ -395,7 +417,7 @@ def api_posts():
                 platform_config
             )
 
-            # Create post object with actual title
+            # Create post object with filter metadata
             post = {
                 'id': post_uuid,
                 'title': post_data.get('title', 'Untitled'),
@@ -409,12 +431,16 @@ def api_posts():
                 'source': post_data.get('source', ''),
                 'source_display': source_display,
                 'tags': post_data.get('tags', []),
-                'external_url': post_data.get('url', '')
+                'external_url': post_data.get('url', ''),
+                # Add filter metadata
+                'filter_score': post_data.get('_filter_score', 0.5),
+                'filter_categories': post_data.get('_filter_categories', []),
+                'filter_tags': post_data.get('_filter_tags', [])
             }
             posts.append(post)
 
-        # Sort by timestamp (newest first)
+        # Sort by filter score (highest first), then timestamp
-        posts.sort(key=lambda x: x['timestamp'], reverse=True)
+        posts.sort(key=lambda x: (x['filter_score'], x['timestamp']), reverse=True)
         
         # Calculate pagination
         total_posts = len(posts)
@@ -1090,17 +1116,18 @@ def settings_filters():
 
     current_filter = user_settings.get('filter_set', 'no_filter')
 
-    # Load available filter sets
+    # Load available filter sets from FilterEngine
-    filter_sets = {}
+    filter_sets_list = []
-    try:
+    for filterset_name in filter_engine.get_available_filtersets():
-        with open('filtersets.json', 'r') as f:
+        description = filter_engine.get_filterset_description(filterset_name)
-            filter_sets = json.load(f)
+        filter_sets_list.append({
-    except:
+            'name': filterset_name,
-        filter_sets = {}
+            'description': description or f'{filterset_name} filter'
+        })
 
     return render_template('settings_filters.html',
                          user=current_user,
-                         filter_sets=filter_sets,
+                         filter_sets=filter_sets_list,
                          current_filter=current_filter)