Skip to content

CHANGELOG.md

Latest commit

 

History

History
54 lines (47 loc) · 2.94 KB

CHANGELOG.md

File metadata and controls

54 lines (47 loc) · 2.94 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[2.0.0a1] - 2026-05-18

Added

  • Complete architectural redesign as hierarchical pest/disease modeling framework
  • Three-layer model hierarchy: Weather -> Agronomic -> Disease
  • PestModelBase ABC with UUID-based model registry and entry-point discovery
  • WeatherProvider ABC with standardized field names
  • NOAA Climate Data Online weather provider
  • GeoID resolver for AgStack Asset Registry integration
  • Threat catalog: 19 crops, 54 crop-threat pairings ported from OpenAgri (8 seeded, expanding)
  • Layer 1 models: GDD (multi-base), Chill Hours (Utah + 32-45), Chill Portions, Accumulated Precipitation
  • Layer 2 models: Leaf Wetness (CART + RH lookup), VPD, Humidity Streak, Phenological Gating
  • Layer 3 models: Fuzzy Mamdani risk engine, Rule-based risk, Powdery Mildew Grape (UC IPM)
  • FastAPI REST API with model discovery, calculation endpoints, crop/threat catalog
  • MCP server (Model Context Protocol) mounted at /mcp for AI agent integration
  • PostgreSQL backend for weather caching and model run audit
  • Docker Compose deployment (Dockerfile, compose.yaml, entrypoint.sh)
  • Sphinx auto-generated documentation
  • FATFD quality harness (.audit/ directory with state, knowledge, guardrails, validator)
  • GitHub Actions CI/CD: 4 quality gates (lint, typecheck, tests, catalog/FATFD integrity)
  • GitHub Actions release pipeline (PyPI wheel, GHCR Docker image, GitHub Pages docs)
  • CODEOWNERS: OpenAgri maintainers as required reviewers for disease models and threat catalog
  • 18 passing unit tests with synthetic weather fixtures
  • Documentation suite:
    • User Manual (library, REST API, and MCP usage with code examples)
    • MCP Server Guide (installation, agent configuration, tool/resource reference)
    • Executive Summary (vision, stakeholders, coverage, community strategy)
    • Design Architecture (three-layer hierarchy, ABCs, CI/CD, FATFD)
    • OpenAgri Migration Guide (evolution narrative, maintainer roles, CI/CD governance)
    • Contributor Guide (adding models, threats, providers, PR process, CODEOWNERS)
    • Deployment Guide (Docker, pip, environment variables)

Changed

  • Package renamed from agralogics to agstack-pnd
  • License changed from GPL-3.0 to Apache-2.0
  • Build system migrated from setup.py to pyproject.toml
  • Weather provider decoupled from proprietary Agralogics API
  • Repository description and topics updated on GitHub

Removed

  • Legacy agralogics package structure
  • Hard dependency on Agralogics Weather Server API

Acknowledgments

  • OpenAgri (Horizon Europe): Fuzzy Mamdani engine, threat catalog, leaf wetness estimation. Stefan Drobic (VizLore), Nikos Kalatzis (Green Supply Chain), Agrobit S.r.l.
  • agstack/pest-models: Original GDD, chill, leaf wetness, powdery mildew models