This directory contains the core implementation of ThemisDB's multi-model database system.
- acceleration/ - GPU and hardware acceleration implementations (CUDA, Vulkan)
- api/ - HTTP API server implementation
- auth/ - Authentication and authorization components (JWT, RBAC)
- cache/ - Semantic caching and query result caching
- cdc/ - Change Data Capture (CDC) and changefeed implementation
- content/ - Content management, ingestion, and processing pipelines
- exporters/ - Data export functionality (JSONL, LLM formats)
- geo/ - Geospatial query processing and indexing
- governance/ - Policy engine and compliance governance
- importers/ - Data import functionality (PostgreSQL, etc.)
- index/ - Index implementations (vector, graph, adaptive, secondary)
- llm/ - LLM interaction storage and chain-of-thought features
- maintenance/ - Centralized database maintenance orchestration (cron scheduling, windows, health reporting)
- plugins/ - Plugin system infrastructure
- process/ - BPMN 2.0/EPK/VCC-VPB process model management, LLM descriptors, Graph-RAG
- query/ - AQL query parser, optimizer, and execution engine
- security/ - Encryption, key management, and PKI integration
- server/ - Main server components and API handlers
- sharding/ - Horizontal scaling and sharding implementation
- storage/ - RocksDB wrapper and storage layer abstractions
- timeseries/ - Time series data management and compression
- transaction/ - SAGA pattern and transaction management
- utils/ - Utility functions and shared components
See the main README.md for build instructions.
For detailed architecture documentation, see:
ThemisDB has 43 Production-grade modules, 4 Beta modules, and 3 Alpha modules across its source tree. Core data paths are production-ready while selected emerging modules are still hardening.
| Tier | Count | Modules |
|---|---|---|
| Production | 43 | 43 of 50 modules |
| Beta | 4 | chimera, gpu, process, sharding |
| Alpha | 3 | ethics_ai, onnx_clip, rpc_grpc |
| Module | Status | Description |
|---|---|---|
| acceleration | 🟢 Production | GPU/hardware acceleration (CUDA, Vulkan) |
| analytics | 🟢 Production | Analytical query pipeline |
| api | 🟢 Production | HTTP API server implementation |
| aql | 🟢 Production | AQL language engine, multi-paradigm queries |
| auth | 🟢 Production | JWT, RBAC, enterprise SSO/MFA |
| base | 🟢 Production | Foundational abstractions |
| cache | 🟢 Production | Semantic and query result caching |
| cdc | 🟢 Production | Change Data Capture and changefeeds |
| chimera | 🟡 Beta | Hybrid multi-model benchmark adapters (simulation complete, production driver wiring pending) |
| config | 🟢 Production | Configuration management |
| content | 🟢 Production | Content ingestion and processing pipelines |
| core | 🟢 Production | Core database runtime |
| exporters | 🟢 Production | Data export (JSONL, LLM formats) |
| ethics_ai | 🔴 Alpha | Ethical reasoning plugin skeleton and discourse engine |
| geo | 🟢 Production | Geospatial query processing and indexing |
| governance | 🟢 Production | Policy engine and compliance governance |
| gpu | 🟡 Beta | GPU compute integration and backend hardening |
| graph | 🟢 Production | Property graph queries and traversal |
| importers | 🟢 Production | Data import (PostgreSQL, etc.) |
| index | 🟢 Production | HNSW, R-tree, adaptive indexing |
| ingestion | 🟢 Production | Data ingestion pipeline |
| llm | 🟢 Production | LLM interaction storage and chain-of-thought |
| maintenance | 🟢 Production | Centralized DB maintenance orchestration |
| metadata | 🟢 Production | Metadata management and catalog |
| network | 🟢 Production | Network layer and peer communication |
| observability | 🟢 Production | Metrics, tracing, and logging infrastructure |
| onnx_clip | 🔴 Alpha | ONNX CLIP image-analysis backend plugin |
| performance | 🟢 Production | Benchmarking and performance optimization |
| plugins | 🟢 Production | Plugin system infrastructure |
| process | 🟡 Beta | BPMN/EPK/VCC-VPB process modeling, Graph-RAG |
| prompt_engineering | 🟢 Production | LLM prompt management |
| query | 🟢 Production | AQL optimizer, cost-based planner, execution engine |
| rag | 🟢 Production | Retrieval-Augmented Generation pipeline |
| replication | 🟢 Production | Raft-based replication |
| rpc_grpc | 🔴 Alpha | Pluggable gRPC RPC server module (mTLS-capable) |
| scheduler | 🟢 Production | Task and job scheduling |
| search | 🟢 Production | Full-text and hybrid search |
| security | 🟢 Production | Encryption, key management, PKI integration |
| server | 🟢 Production | Main server components and API handlers |
| sharding | 🟡 Beta | Horizontal scaling and sharding |
| storage | 🟢 Production | RocksDB wrapper, MVCC, backup/recovery |
| temporal | 🟢 Production | Temporal and bitemporal data support |
| themis | 🟢 Production | Core ThemisDB orchestration layer |
| timeseries | 🟢 Production | Time series data management and compression |
| training | 🟢 Production | ML model training integration |
| transaction | 🟢 Production | SAGA pattern and distributed transactions |
| updates | 🟢 Production | Schema and data update management |
| user_storage_encrypted | 🟢 Production | Encrypted user storage backend (gocryptfs + key rotation) |
| utils | 🟢 Production | Shared utility functions and helpers |
| voice | 🟢 Production | Voice query interface |
Source of truth: The
Maturity Levelfield in each source file header is the authoritative status for that file. The table above reflects the overall module status derived from those headers.
| Level | Meaning |
|---|---|
| 🟢 Production | Feature-complete, thoroughly tested, used in production deployments. API is stable. |
| 🟡 Beta | Core functionality works and is tested. May have rough edges; API may change in minors. |
| 🔴 Alpha | Under active development. Functionality may be incomplete. Not recommended for production. |
See ROADMAP.md for the development timeline and graduation milestones for each module.