Merge remote-tracking branch 'jaychao/steven_update' into merge_pr

Author: bchou9

.gitignore

@@ -35,6 +35,7 @@ coverage/
 *.coverage
 backend/.coverage
 node_modules
+backend/qdrant_storage/
 frontend/build/
 
 # Exclude all .env files

backend/NEXT_STEPS.md

@@ -0,0 +1 @@
+

backend/SETUP_VECTOR_SEARCH.md

@@ -0,0 +1,66 @@
+# Vector Search Setup Guide
+
+## Quick Start
+
+### Option 1: Using Docker Compose (Recommended)
+
+```bash
+cd backend
+
+# Start Qdrant and Redis
+docker-compose up -d qdrant redis
+
+# Verify Qdrant is running
+curl http://localhost:6333/healthz
+# Should return: {"title":"healthz","version":"1.x.x"}
+
+# Install Python dependencies (if not already done)
+pip install -r requirements.txt
+
+# Run backend
+python app.py
+
+# Seperate terminal 
+python worker/embedding_service.py 
+```
+
+### Option 2: Local Qdrant Installation
+
+```bash
+# macOS with Homebrew
+brew install qdrant
+
+# Or using Docker standalone
+docker run -p 6333:6333 -p 6334:6334 \
+  -v $(pwd)/qdrant_storage:/qdrant/storage \
+  qdrant/qdrant
+
+# Install Python dependencies
+cd backend
+pip install -r requirements.txt
+
+# Run backend
+python app.py
+
+# Seperate terminal 
+python worker/embedding_service.py 
+```
+
+---
+
+## 🔧 Configuration
+
+The following environment variables can be set in `config.py`:
+
+```bash
+# Qdrant Vector Database
+QDRANT_HOST=localhost
+QDRANT_PORT=6333
+QDRANT_GRPC_PORT=6334
+QDRANT_COLLECTION_NAME=rescanvas_embeddings
+
+# These are already in your config:
+# REDIS_HOST=localhost
+# REDIS_PORT=6379
+# MONGO_ATLAS_URI=...
+```

backend/app.py

@@ -39,6 +39,9 @@ def filter(self, record):
 
 app = Flask(__name__)
 
+# Allow large request bodies for thumbnail uploads (up to 20MB)
+app.config['MAX_CONTENT_LENGTH'] = 20 * 1024 * 1024
+
 # Initialize rate limiting BEFORE importing routes (routes use limiter decorators)
 from middleware.rate_limit import init_limiter, rate_limit_error_handler
 limiter = init_limiter(app)
@@ -57,6 +60,7 @@ def filter(self, record):
 from routes.analytics import analytics_bp
 from routes.export import export_bp
 from routes.ai_assistant import ai_assistant_bp
+from routes.search_ai import search_ai_bp
 from services.db import redis_client
 from services.canvas_counter import get_canvas_draw_count
 from services.graphql_service import commit_transaction_via_graphql
@@ -234,6 +238,7 @@ def handle_all_exceptions(e):
 app.register_blueprint(users_v1_bp)
 app.register_blueprint(stamps_bp, url_prefix='/api')
 app.register_blueprint(templates_v1_bp)
+app.register_blueprint(search_ai_bp)
 
 # Frontend serving must be last to avoid route conflicts
 app.register_blueprint(frontend_bp)

backend/config.py

@@ -46,6 +46,13 @@
 REDIS_HOST = os.getenv("REDIS_HOST", "localhost")
 REDIS_PORT = int(os.getenv("REDIS_PORT", "6379"))
 
+# Qdrant Vector Database Configuration
+QDRANT_HOST = os.getenv("QDRANT_HOST", "localhost")
+QDRANT_PORT = int(os.getenv("QDRANT_PORT", "6333"))
+QDRANT_COLLECTION_NAME = os.getenv("QDRANT_COLLECTION_NAME", "rescanvas_embeddings")
+EMBEDDING_DIMENSION = 512  # OpenCLIP ViT-B-32 output dimension
+QDRANT_GRPC_PORT = int(os.getenv("QDRANT_GRPC_PORT", "6334"))
+
 # Rate Limiting Configuration
 RATE_LIMIT_STORAGE_URI = f"redis://{REDIS_HOST}:{REDIS_PORT}"
 RATE_LIMIT_ENABLED = os.getenv("RATE_LIMIT_ENABLED", "True") == "True"

backend/docker-compose.yml

@@ -18,6 +18,25 @@ services:
     networks:
       - rescanvas-network
 
+  qdrant:
+    image: qdrant/qdrant:latest
+    container_name: rescanvas-qdrant
+    ports:
+      - "6333:6333"  # HTTP API
+      - "6334:6334"  # gRPC API
+    volumes:
+      - qdrant_data:/qdrant/storage
+    environment:
+      - QDRANT__SERVICE__GRPC_PORT=6334
+    restart: unless-stopped
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:6333/healthz"]
+      interval: 10s
+      timeout: 5s
+      retries: 5
+    networks:
+      - rescanvas-network
+
   backend:
     build:
       context: .
@@ -34,6 +53,9 @@ services:
       - RATE_LIMIT_STORAGE_URI=redis://redis:6379
       - REDIS_HOST=redis
       - REDIS_PORT=6379
+      - QDRANT_HOST=qdrant
+      - QDRANT_PORT=6333
+      - QDRANT_GRPC_PORT=6334
       - JWT_SECRET=${JWT_SECRET:-dev-insecure-change-me}
       - OPENAI_API_KEY=${OPENAI_API_KEY}
       - ANALYTICS_ENABLED=${ANALYTICS_ENABLED:-True}
@@ -46,6 +68,8 @@ services:
     depends_on:
       redis:
         condition: service_healthy
+      qdrant:
+        condition: service_healthy
     restart: unless-stopped
     healthcheck:
       test: ["CMD", "curl", "-f", "http://localhost:10010/api/analytics/health"]
@@ -59,6 +83,8 @@ services:
 volumes:
   redis_data:
     driver: local
+  qdrant_data:
+    driver: local
 
 networks:
   rescanvas-network:

backend/requirements.txt

@@ -51,7 +51,7 @@ python-engineio==4.12.3
 python-socketio==5.14.1
 redis==6.2.0
 requests==2.32.4
-resilient-python-cache==0.1.1
+# resilient-python-cache==0.1.1
 rich==13.9.4
 simple-websocket==1.1.0
 simplejson==3.19.3
@@ -64,3 +64,9 @@ websockets==10.4
 Werkzeug==3.1.3
 wrapt==2.0.0
 wsproto==1.2.0
+# AI/ML dependencies for semantic search
+torch>=2.0.0
+open_clip_torch>=2.20.0
+pillow>=10.0.0
+qdrant-client>=1.7.0
+numpy>=1.24.0

backend/routes/rooms.py

@@ -3277,3 +3277,82 @@ def notification_preferences():
     except Exception:
         return jsonify({"status":"error","message":"Failed to persist preferences"}), 500
     return jsonify({"status":"ok","preferences": clean})
+
+
+@rooms_bp.route("/rooms/<roomId>/thumbnail", methods=["POST"])
+@require_auth
+@require_room_access(room_id_param='roomId')
+def upload_room_thumbnail(roomId):
+    try:
+        data = request.get_json()
+        if not data:
+            return jsonify({"error": "Request body required"}), 400
+        
+        thumbnail_data = data.get('thumbnail')
+        if not thumbnail_data:
+            return jsonify({"error": "thumbnail field required"}), 400
+        
+        # Strip data URL prefix if present
+        # Format: data:image/png;base64,iVBORw0KG...
+        if thumbnail_data.startswith('data:'):
+            if ',' in thumbnail_data:
+                thumbnail_data = thumbnail_data.split(',', 1)[1]
+            else:
+                return jsonify({"error": "Invalid data URL format"}), 400
+        
+        # Decode base64 to binary
+        import base64
+        try:
+            thumbnail_bytes = base64.b64decode(thumbnail_data)
+        except Exception as e:
+            logger.error(f"Failed to decode thumbnail base64 for room {roomId}: {e}")
+            return jsonify({"error": "Invalid base64 encoding"}), 400
+        
+        # Validate minimum size (at least 100 bytes for a valid image)
+        if len(thumbnail_bytes) < 100:
+            return jsonify({"error": "Thumbnail too small, likely invalid"}), 400
+        
+        # Validate maximum size (10MB limit)
+        if len(thumbnail_bytes) > 10 * 1024 * 1024:
+            return jsonify({"error": "Thumbnail too large (max 10MB)"}), 400
+        
+        # Optional: Validate it's actually a PNG/JPEG using magic bytes
+        # PNG magic bytes: 89 50 4E 47
+        # JPEG magic bytes: FF D8 FF
+        is_png = thumbnail_bytes[:4] == b'\x89PNG'
+        is_jpeg = thumbnail_bytes[:3] == b'\xff\xd8\xff'
+        
+        if not (is_png or is_jpeg):
+            logger.warning(f"Thumbnail for room {roomId} doesn't appear to be PNG or JPEG")
+            # Don't reject, just log warning
+        
+        # Store thumbnail in room document
+        updated_at = datetime.utcnow()
+        result = rooms_coll.update_one(
+            {'_id': ObjectId(roomId)},
+            {
+                '$set': {
+                    'thumbnail': thumbnail_bytes,  # Binary data
+                    'thumbnailUpdatedAt': updated_at,
+                    'updatedAt': updated_at  # Also update room's main timestamp
+                }
+            }
+        )
+        
+        if result.matched_count == 0:
+            return jsonify({"error": "Room not found"}), 404
+        
+        logger.info(f"Stored thumbnail for room {roomId}: {len(thumbnail_bytes)} bytes "
+                   f"(format: {'PNG' if is_png else 'JPEG' if is_jpeg else 'unknown'})")
+        
+        return jsonify({
+            "status": "success",
+            "roomId": roomId,
+            "thumbnailSize": len(thumbnail_bytes),
+            "format": "PNG" if is_png else "JPEG" if is_jpeg else "unknown",
+            "updatedAt": updated_at.isoformat()
+        }), 200
+        
+    except Exception as e:
+        logger.exception(f"Failed to upload thumbnail for room {roomId}: {e}")
+        return jsonify({"error": "Internal server error", "details": str(e)}), 500