DataJoint 2.1 Released — PostgreSQL Support, Improved Diagrams, and More

[dimitri-yatsenko]

DataJoint 2.1 is now available on PyPI (pip install datajoint).

Highlights

PostgreSQL backend support

DataJoint now supports PostgreSQL as an alternative to MySQL. Set dj.config['database.backend'] = 'postgresql' to connect to PostgreSQL databases. (#1338, #1339, #1340)

Diagram improvements

• collapse() method for high-level pipeline views
• Mermaid output format via output='mermaid'
• Schema grouping with module labels
• Direction control (direction='LR' or 'TB')
• Default direction changed from TB to LR

Singleton tables

Support for tables with empty primary keys. (#1341)

Performance and packaging

• Lazy-load deepdiff and tqdm in autopopulate for faster imports (refactor: code organization and lazy imports for better modularity #1349)
• Switched from setuptools to hatchling (fix: use hatchling to build package #1358)

Bug fixes (2.1.1)

• Atomic job reservation to prevent race conditions (fix: Atomic job reservation to prevent race condition #1399)
• Correct Part table names in diagrams (fix: Strip module prefix correctly for Part tables in diagrams #1392)
• Hide SQL comments from .preview() output (fix: Hide comments from table preview display #1393)
• Removed size_on_disk property — use database-native tools (Remove size_on_disk from Table and Schema #1395)

Links

• Release 2.1.0
• Release 2.1.1
• Full changelog
• Documentation