v0.2.2

What's Changed

• feat(rapid):revise rapid classifier impl by @RingsC in #676
• feat(ML):revise ml function by @ShannonBase in #678
• feat(ML):revise explanation of regression by @ShannonBase in #679
• feat(ML): revise explain table impl by @RingsC in #680
• feat(ML):revise ml predicate table by @RingsC in #681
• feat(ML):revise ml active impl by @RingsC in #684
• feat(javascript):to support more types by @RingsC in #685
• feat(fix):hot fix mtr result by @RingsC in #687
• feat(jerryscript): to enable execute sql by @RingsC in #688
• feat(javascript): to support type info in execution by @RingsC in #691
• feat(fix):fix none-return statement by @RingsC in #692
• feat(fix):fix model detector by @RingsC in #693

Full Changelog: ShannonBase-V0.2.1...ShannonBase-v0.2.2

v0.2.1

What's Changed

Features
Vector Support: Fixed and refined vector type support, including improvements to vector scan iterator implementation (#656, #657)
Machine Learning & AI:
Added schema metadata retrieval support for ML operations (#660)
Enabled PII (Personally Identifiable Information) detector functionality (#666, #667)
Revised schema metadata retrieval implementation (#671)
Reviewed and refined ML algorithms (#673)
Recovery: Refined recovery implementation and fixed recovery-related crashes (#659, #670)
JavaScript Stored Procedures: Fixed missing files for JavaScript stored procedure support (#663)
Full Changelog: ShannonBase-v0.2...ShannonBase-V0.2.1

v0.2

What's Changed

1: Upgrade the mysql core from 8.4.4 to 8.4.8
2: Support Hypergraph optimization
3: Support Recovery from instance crash or reboot.

v0.1

Rapid Columnar Engine (IMCS)

• Native In-Memory Column Store for HTAP workloads

• Intelligent workload routing between InnoDB and Rapid using cost-based and ML-based algorithms

• MVCC support through version linking

• Automatic synchronous propagation of InnoDB changes via Redo logs

Multi-Modal Data Support

• Structured, semi-structured, and unstructured data types

• GIS, JSON, and Vector data types

Machine Learning Capabilities

• Native support for LightGBM/XGBoost integration

• SQL-based ML operations: ml_train, ml_predict_row, ml_model_import

• Classification, regression, recommendation, and anomaly detection

• Pre-trained model import support

Vector & Embedding Support

• Native vector data type

• Zero data movement for ML/AI workloads**

• Optimized for RAG and embedding algorithms

• Seamless SQL integration

Multilingual Engine

• Lightweight JerryScript JavaScript engine integration

• Write stored procedures in SQL or JavaScript

ShannonBase v0.1 delivers a unified infrastructure for data processing and Generative AI, extending MySQL with native AI capabilities while maintaining ease of use for data scientists and developers.