What is New in apsimNGpy 1.0.0
apsimNGpy 1.0.0 represents a major milestone in the development of the framework, transitioning from an experimental research tool to a stable, production-ready release. This version consolidates years of development and introduces several key improvements in performance, usability, reproducibility, and analytical capability.
1. Stable, Reproducible Release
First official 1.0.0 version, signifying a stable API and backward compatibility guarantee for future minor releases.
2. Core Engine Improvements
Refactored multiprocessing engine for robust, scalable execution across multiple CPU cores, including safer handling of parallel APSIM runs on Windows.
Improved failure reporting and retry mechanisms with configurable policies (e.g.,
tenacity-based retries), reducing silent errors in large batch jobs.Improve job submissions allowing edits to be submitted simultaneously.
3. Expanded Sensitivity & Uncertainty Analysis
Updated Sobol sampling with configurable skip values for improved space-filling design.
Clean handling of calc_second_order options with consistent propagation between sampling and analysis layers.
Support for additional SALib methods with stable default parameterization.
Sensitivity analysis workflows fully compatible across all os platforms.
4. Improved Database & Output Management
Schema-hash table naming to avoid SQLite collisions in parallel executions.
Stable persistence layer with: * deterministic table identifiers * execution and process metadata * large result handling with chunked writes.
Cleaner error handling for results writes under heavy parallel loads.
5. Workflow & Developer Quality-of-Life
First modules test using .bat scripts
Support for locking APSIM versions to a specific project.
6. Fixes & Stability Enhancements
Resolution of common parallel SQLite locking issues under heavy batch throughput.
Deterministic hashing for table identifiers even in multiprocessing contexts.
Guidance and preflight validation for schema drift, unsupported data types, and mixed index/column structures.
Better error reporting for model editing callbacks and APSIM parameter sets.
Summary
apsimNGpy 1.0.0 delivers:
A stable, reproducible foundation for agri-environmental modeling workflows
Scalability, reliability for large batch, single and multi-objective experiments
Better integration of APSIM with decision support, sensitivity, and spatial optimization routines
An enduring API that is resilient and robust under a wide range of uncertainties
This release establishes a platform for future enhancements while remaining reliable for academic and applied research in productivity, environmental impacts, and landscape planning.