EazyDataFix v2.0.0

EazyDataFix 2.0.0 Release Notes

Version 2.0.0 adds versioned preparation recipes, explicit column conversion,
preview/apply review plans, transformation evidence, reusable ML inference
artifacts and recipe-enabled command-line workflows. Existing v1 APIs retain
their defaults and return types.

Install with pip install --upgrade eazydatafix==2.0.0.
For YAML recipes use pip install 'eazydatafix[yaml]==2.0.0'.

New features

• Versioned JSON/YAML recipes with strict schema validation and no executable config.
• Declared string, integer, number, boolean, date and category conversions.
• Explicit numeric separators, currency tokens, date formats and category mappings.
• Preserve, reject or quarantine invalid conversions with source-value evidence.
• Preview, select stages and apply with source, recipe and output integrity checks.
• Transformation counts, dtype changes, bounded examples and optional detailed audits.
• Common reports and gated data exports across Analysis Ready, ML Ready and Power BI Ready.
• Recipe-bound ML inference artifacts using training-fitted preprocessing parameters.
• Multi-table Power BI recipes with existing key, relationship and date-table validation.
• CLI recipe previews, named tables, exports and readiness exit codes.
• CSV/Excel token preservation and duplicate-heading validation before inference.

Existing profiling, assessment, cleaning, contracts, EDA, reporting and v1 readiness
APIs remain available. See MIGRATION_GUIDE.md for adoption guidance.

See the recipe guide for runnable examples, compatibility
and scope. Browser Studio integration, streaming, SQL connectors and persistent
cross-process approval plans remain future work. Review plans currently live in
the Python process; recipes and ML artifacts can be saved and reloaded.