fix(server): support dedicated backend data structures and serialization logic for vector index.

.gitignore

@@ -91,6 +91,8 @@ hs_err_pid*
 hugegraph-server/hugegraph-dist/docker/data/
 
 # AI-IDE prompt files (We only keep AGENTS.md, other files could soft-linked it when needed)
+# Serena MCP memories
+.serena/
 # Claude Projects
 CLAUDE.md
 CLAUDE_*.md

.licenserc.yaml

@@ -59,6 +59,7 @@ header: # `header` section is configurations for source codes license header.
     - 'LICENSE'
     - 'NOTICE'
     - 'DISCLAIMER'
+    - '.serena/**'
     - '**/*.versionsBackup'
     - '**/*.versionsBackup'
     - '**/*.proto'

.serena/.gitignore

@@ -0,0 +1 @@
+/cache

.serena/memories/architecture_and_modules.md

@@ -0,0 +1,99 @@
+# Architecture and Module Structure
+
+## Three-Tier Architecture
+
+### 1. Client Layer
+- Gremlin/Cypher query interfaces
+- REST API endpoints
+- Multiple client language bindings
+
+### 2. Server Layer (hugegraph-server)
+- **REST API Layer** (hugegraph-api): GraphAPI, SchemaAPI, GremlinAPI, CypherAPI, AuthAPI
+- **Graph Engine Layer** (hugegraph-core): Schema management, traversal optimization, task scheduling
+- **Backend Interface**: Abstraction over storage backends
+
+### 3. Storage Layer
+- Pluggable backend implementations
+- Each backend extends `hugegraph-core` abstractions
+- Implements `BackendStore` interface
+
+## Multi-Module Structure
+
+The project consists of 7 main modules:
+
+### 1. hugegraph-server (13 submodules)
+Core graph engine, REST APIs, and backend implementations:
+- `hugegraph-core` - Core graph engine and abstractions
+- `hugegraph-api` - REST API implementations (includes OpenCypher in `opencypher/`)
+- `hugegraph-dist` - Distribution packaging and scripts
+- `hugegraph-test` - Test suites (unit, core, API, TinkerPop)
+- `hugegraph-example` - Example code
+- Backend implementations:
+  - `hugegraph-rocksdb` (default)
+  - `hugegraph-hstore` (distributed)
+  - `hugegraph-hbase`
+  - `hugegraph-mysql`
+  - `hugegraph-postgresql`
+  - `hugegraph-cassandra`
+  - `hugegraph-scylladb`
+  - `hugegraph-palo`
+
+### 2. hugegraph-pd (8 submodules)
+Placement Driver for distributed deployments (meta server):
+- `hg-pd-core` - Core PD logic
+- `hg-pd-service` - PD service implementation
+- `hg-pd-client` - Client library
+- `hg-pd-common` - Shared utilities
+- `hg-pd-grpc` - gRPC protocol definitions (auto-generated)
+- `hg-pd-cli` - Command line interface
+- `hg-pd-dist` - Distribution packaging
+- `hg-pd-test` - Test suite
+
+### 3. hugegraph-store (9 submodules)
+Distributed storage backend with RocksDB and Raft:
+- `hg-store-core` - Core storage logic
+- `hg-store-node` - Storage node implementation
+- `hg-store-client` - Client library
+- `hg-store-common` - Shared utilities
+- `hg-store-grpc` - gRPC protocol definitions (auto-generated)
+- `hg-store-rocksdb` - RocksDB integration
+- `hg-store-cli` - Command line interface
+- `hg-store-dist` - Distribution packaging
+- `hg-store-test` - Test suite
+
+### 4. hugegraph-commons
+Shared utilities across modules:
+- Locks and concurrency utilities
+- Configuration management
+- RPC framework components
+
+### 5. hugegraph-struct
+Data structure definitions shared between modules.
+**Important**: Must be built before PD and Store modules.
+
+### 6. install-dist
+Distribution packaging and release management:
+- License and NOTICE files
+- Dependency management scripts
+- Release documentation
+
+### 7. hugegraph-cluster-test
+Cluster integration tests for distributed deployments
+
+## Cross-Module Dependencies
+
+```
+hugegraph-commons → (shared by all modules)
+hugegraph-struct → hugegraph-pd + hugegraph-store
+hugegraph-core → (extended by all backend implementations)
+```
+
+## Distributed Architecture (Optional)
+
+For production distributed deployments:
+- **hugegraph-pd**: Service discovery, partition management, metadata
+- **hugegraph-store**: Distributed storage with Raft (3+ nodes)
+- **hugegraph-server**: Multiple server instances (3+)
+- Communication: All use gRPC with Protocol Buffers
+
+**Status**: Distributed components (PD + Store) are in BETA

.serena/memories/code_style_and_conventions.md

@@ -0,0 +1,92 @@
+# Code Style and Conventions
+
+## Code Style Configuration
+- **Import**: Use `hugegraph-style.xml` in your IDE (IntelliJ IDEA recommended)
+- **EditorConfig**: `.editorconfig` file defines style rules (validated in CI)
+- **Checkstyle**: `style/checkstyle.xml` defines additional rules
+
+## Core Style Rules (from .editorconfig)
+
+### General
+- Charset: UTF-8
+- End of line: LF (Unix-style)
+- Insert final newline: true
+- Max line length: 100 characters (120 for XML)
+- Visual guides at column 100
+
+### Java Files
+- Indent: 4 spaces (not tabs)
+- Continuation indent: 8 spaces
+- Wrap on typing: true
+- Wrap long lines: true
+
+### Import Organization
+```
+$*
+|
+java.**
+|
+javax.**
+|
+org.**
+|
+com.**
+|
+*
+```
+- Class count to use import on demand: 100
+- Names count to use import on demand: 100
+
+### Formatting Rules
+- Line comments not at first column
+- Align multiline: chained methods, parameters in calls, binary operations, assignments, ternary, throws, extends, array initializers
+- Wrapping: normal (wrap if necessary)
+- Brace forcing:
+  - if: if_multiline
+  - do-while: always
+  - while: if_multiline
+  - for: if_multiline
+- Enum constants: split_into_lines
+
+### Blank Lines
+- Max blank lines in declarations: 1
+- Max blank lines in code: 1
+- Blank lines between package declaration and header: 1
+- Blank lines before right brace: 1
+- Blank lines around class: 1
+- Blank lines after class header: 1
+
+### Documentation
+- Add `<p>` tag on empty lines: true
+- Do not wrap if one line: true
+- Align multiline annotation parameters: true
+
+### XML Files
+- Indent: 4 spaces
+- Max line length: 120
+- Text wrap: off
+- Space inside empty tag: true
+
+### Maven
+- Compiler source/target: Java 11
+- Max compiler errors: 500
+- Compiler args: `-Xlint:unchecked`
+- Source encoding: UTF-8
+
+## Lombok Usage
+- Version: 1.18.30
+- Scope: provided
+- Optional: true
+
+## License Headers
+- All source files MUST include Apache Software License header
+- Validated by apache-rat-plugin and skywalking-eyes
+- Exclusions defined in pom.xml (line 171-221)
+- gRPC generated code excluded from license check
+
+## Naming Conventions
+- Package names: lowercase, dot-separated (e.g., org.apache.hugegraph)
+- Class names: PascalCase
+- Method names: camelCase
+- Constants: UPPER_SNAKE_CASE
+- Variables: camelCase

.serena/memories/ecosystem_and_related_projects.md

@@ -0,0 +1,63 @@
+# HugeGraph Ecosystem and Related Projects
+
+## Core Repository (This Project)
+**Repository**: apache/hugegraph (server)
+**Purpose**: Core graph database engine (OLTP)
+
+## Related Repositories
+
+### 1. hugegraph-toolchain
+**Repository**: https://github.com/apache/hugegraph-toolchain
+**Components**:
+- **hugegraph-loader**: Bulk data loading tool
+- **hugegraph-hubble**: Web-based visualization dashboard
+- **hugegraph-tools**: Command-line utilities
+- **hugegraph-client**: Java client SDK
+
+### 2. hugegraph-computer
+**Repository**: https://github.com/apache/hugegraph-computer
+**Purpose**: Distributed graph computing framework (OLAP)
+**Features**: PageRank, Connected Components, Shortest Path, Community Detection
+
+### 3. hugegraph-ai
+**Repository**: https://github.com/apache/incubator-hugegraph-ai
+**Purpose**: Graph AI, LLM, and Knowledge Graph integration
+**Features**: Graph-enhanced LLM, KG construction, Graph RAG, NL to Gremlin/Cypher
+
+### 4. hugegraph-website
+**Repository**: https://github.com/apache/hugegraph-doc
+**Purpose**: Official documentation and website
+**URL**: https://hugegraph.apache.org/
+
+## Integration Points
+
+### Data Pipeline
+```
+Data Sources → hugegraph-loader → hugegraph-server
+                                        ↓
+                    ┌───────────────────┼───────────────────┐
+                    ↓                   ↓                   ↓
+            hugegraph-hubble    hugegraph-computer   hugegraph-ai
+            (Visualization)      (Analytics)         (AI/ML)
+```
+
+## External Integrations
+
+### Big Data Platforms
+- Apache Flink, Apache Spark, HDFS
+
+### Storage Backends
+- RocksDB (default), HBase, Cassandra, ScyllaDB, MySQL, PostgreSQL
+
+### Query Languages
+- Gremlin (Apache TinkerPop), Cypher (OpenCypher), REST API
+
+## Version Compatibility
+- Server: 1.7.0
+- TinkerPop: 3.5.1
+- Java: 11+ required
+
+## Use Cases
+- Social networks, Fraud detection, Recommendation systems
+- Knowledge graphs, Network analysis, Supply chain management
+- IT operations, Bioinformatics

.serena/memories/implementation_patterns_and_guidelines.md

@@ -0,0 +1,104 @@
+# Implementation Patterns and Guidelines
+
+## Backend Development
+
+### Backend Architecture Pattern
+- All backends extend abstractions from `hugegraph-server/hugegraph-core`
+- Implement the `BackendStore` interface
+- Each backend is a separate Maven module under `hugegraph-server/`
+- Backend selection configured in `hugegraph.properties` via `backend` property
+
+### Available Backends
+- **RocksDB** (default, embedded): `hugegraph-rocksdb`
+- **HStore** (distributed, production): `hugegraph-hstore`
+- **Legacy** (≤1.5.0): MySQL, PostgreSQL, Cassandra, ScyllaDB, HBase, Palo
+
+### Backend Testing Profiles
+- `memory`: In-memory backend for fast unit tests
+- `rocksdb`: RocksDB for realistic local tests
+- `hbase`: HBase for distributed scenarios
+- `hstore`: HStore for production-like distributed tests
+
+## gRPC Protocol Development
+
+### Protocol Buffer Definitions
+- PD protos: `hugegraph-pd/hg-pd-grpc/src/main/proto/`
+- Store protos: `hugegraph-store/hg-store-grpc/src/main/proto/`
+
+### Code Generation
+When modifying `.proto` files:
+1. Run `mvn clean compile` to regenerate gRPC stubs
+2. Generated Java code goes to `*/grpc/` packages
+3. Output location: `target/generated-sources/protobuf/`
+4. Generated files excluded from Apache RAT checks
+5. All inter-service communication uses gRPC
+
+## Authentication System
+
+### Default State
+- Authentication **disabled by default**
+- Enable via `bin/enable-auth.sh` or configuration
+- **Required for production deployments**
+
+### Implementation Location
+`hugegraph-server/hugegraph-api/src/main/java/org/apache/hugegraph/api/auth/`
+
+### Multi-Level Security Model
+- Users, Groups, Projects, Targets, Access control
+
+## TinkerPop Integration
+
+### Compliance
+- Full Apache TinkerPop 3 implementation
+- Custom optimization strategies
+- Supports both Gremlin and OpenCypher query languages
+
+### Query Language Support
+- **Gremlin**: Native via TinkerPop integration
+- **OpenCypher**: Implementation in `hugegraph-api/opencypher/`
+
+## Testing Patterns
+
+### Test Suite Organization
+- **UnitTestSuite**: Pure unit tests, no external dependencies
+- **CoreTestSuite**: Core functionality tests with backend
+- **ApiTestSuite**: REST API integration tests
+- **StructureStandardTest**: TinkerPop structure compliance
+- **ProcessStandardTest**: TinkerPop process compliance
+
+### Backend Selection in Tests
+Use Maven profiles:
+```bash
+-P core-test,memory      # Fast in-memory
+-P core-test,rocksdb     # Persistent local
+-P api-test,rocksdb      # API with persistent backend
+```
+
+## Distribution and Packaging
+
+### Creating Distribution
+```bash
+mvn clean package -DskipTests
+```
+Output: `install-dist/target/hugegraph-<version>.tar.gz`
+
+## Code Organization
+
+### Package Structure
+```
+org.apache.hugegraph
+├── backend/        # Backend implementations
+├── api/           # REST API endpoints
+├── core/          # Core graph engine
+├── schema/        # Schema definitions
+├── traversal/     # Traversal and query processing
+├── task/          # Background tasks
+├── auth/          # Authentication/authorization
+└── util/          # Utilities
+```
+
+### Module Dependencies
+- Commons is shared by all modules
+- Struct must be built before PD and Store
+- Backend modules depend on core
+- Test module depends on all server modules