v4.3.5 🌈

Data Type Support

Link to bytes support - Added support for link to bytes data type (DEE-299)
Flexible text column handling - Text columns now allow appending data with unknown dtype for better compatibility

🐛 Bug Fixes

deeplake.exists error handling - Fixed exception handling in deeplake.exists to gracefully handle errors instead of throwing exceptions (DEE-300)
Python PNG linking - Resolved Python PNG linking issues

📚 Documentation

Documentation updates - Improved and updated documentation (DEE-36)

🧹 Code Quality

General improvements - Various small fixes and code improvements

v4.3.4

🎉 New Features
PostgreSQL 18 Support

PostgreSQL 18 compatibility - pg_deeplake now supports PostgreSQL 18

Performance Optimizations

Adaptive backoff for worker threads - Implemented an intelligent backoff strategy for idle worker threads to reduce CPU usage
Cgroup-aware concurrency - Hardware concurrency now respects cgroup limits for better container performance

Azure Integration

Azure container name handling - Improved Azure getter with proper container name handling

🐛 Bug Fixes

Build errors resolved - Fixed compilation errors in pg_deeplake
Visualizer fixes - Resolved issues in the visualizer component

🧹 Code Quality

Code cleanup - Improved code organization with proper renames and refactoring
Modular architecture - Continued separation of concerns for better maintainability

🧪 Testing

Enhanced test coverage - Added comprehensive tests for pg_deeplake

v4.3.0 🌈

Deeplake 4.3.0

Deeplake 4.3.0 is a major update bringing many new features to the Deeplake ecosystem.

New Data and Index Types

• Complete revisit of Sequence types to support visual and structured data
• Video type support is now available in Deeplake, supporting MP4 and MKV videos with H264 codec and providing fast random access to video frames
• Indexing for numeric types, enabling fast queries for numeric comparisons in TQL, including IN and BETWEEN operations
• Significant improvements to textual index types, providing faster search without requiring index regeneration

Data Import/Export

• Fully rewritten from_csv function with support for large CSV files
• New to_csv API to export Deeplake datasets/views to CSV format

Python Typing

• Support for specifying Python builtin types when defining dataset schemas
• Support for using Pydantic Models as dataset schemas
• Enriched async operations typing, to support better integration with linters and IDEs.

Improvements and Bug Fixes

• Improved TQL data fetching and linear scan performance for non-indexed columns
• Better memory usage tracking to prevent out-of-memory errors
• Various stability improvements and bug fixes

Compatibility Notice

Deeplake 4.3.0 is backward compatible with datasets created in v4.2.x. However, datasets created or modified with v4.3.0 cannot be opened with v4.2.x versions due to internal format enhancements. We recommend upgrading all environments to v4.3.0 when working with shared datasets.

v3.9.51 🌈

🧭 What's Changed

• Update version for 3.9.51 release (#3073) @activesoull
• [DEE-180] fix dependency resolver issue (#3072) @activesoull

⚙️ Who Contributes

@activesoull

v3.9.50 🌈

🧭 What's Changed

• Updated versions for 3.9.50 release (#3069) @activesoull
• Refactor frame extraction logic and improve frame rate retrieval in L… (#3071) @azat-manukyan

⚙️ Who Contributes

@activesoull and @azat-manukyan

v3.9.46 🌈

🧭 What's Changed

• Base mmsegmentation dataset (#3060) @activesoull
  added a base mmsegmentation dataset class where in case of using mmseg 1.x version you need to inherit from to make integration workflow smoother

• 3.9.46 (#3058) @activesoull

⚙️ Who Contributes

@activesoull

v4.2.14 🌈

• Added Audio type
• Fixed conversion from V3 for Polygon type

v4.2.12 🌈

• Added autocommit.
• Improved parallel ingestion and compaction functionality
• Added zlib compression support for SegmentMask type
• Added embeddings indexing support for float16 and bfloat16
• Added AVG function to TQL
• Improved SAMPLE BY in TQL
• Added images support for labelbox integration

v4.2.8 🌈

• Added from_csv, extended from_parquet to support content bytes
• Added zlib and lz4 sample compression for SegmentMask type

v3.9.45 🌈

🧭 What's Changed

• added sampler setter function (#3057) @activesoull
• Hierarchical namespace (#3056) @activesoull
• Updated versions for 3.9.45 release (#3045) @activesoull

⚙️ Who Contributes

@activesoull