[general]
name=pypopRF
qgisMinimumVersion=3.44
qgisMaximumVersion=4.99
description=A plugin for population prediction and dasymetric mapping using machine learning
version=0.3.8
author=WorldPop SDI Team:
    Borys Nosatiuk (Project Lead),
    Rhorom Priyatikanto,
    Maksym Bondarenko,
    Wenbin Zhang,
    Tom McKeen,
    Elena Vataga
email=b.nosatiuk@soton.ac.uk
about=Create gridded population maps from census counts and aligned raster covariates. Calculations run in an isolated 64-bit Python 3.12 environment managed by the plugin. Qualified targets are Windows AMD64 and Linux glibc x86_64; macOS, ARM and musl are not qualified. Initial engine setup downloads locked Python packages; activation never installs packages into QGIS. Census totals are verified separately for every zone. Model validation describes held-out census zones and does not establish pixel-level accuracy. See the packaged guides for installation, updates and validation limits.
tracker=https://github.com/wpgp/QGIS-pypopRF/issues
repository=https://github.com/wpgp/QGIS-pypopRF
hasProcessingProvider=no
changelog=Version 0.3.8 (2026-10-15):
    * Current QGIS enum paths for Qt5/Qt6 without obsolete fallback references
    * Explicit argv-only engine processes; reject shell wrappers at process boundary
    * Deterministic reviewed-checksum baseline without disabling secret detectors
    * Independent packaged security checks and official QGIS Qt6 checker in CI
Version 0.3.7 (2026-10-15):
    * Isolated hash-verified Python 3.12 engine, no QGIS package injection
    * Independent light Qt5/Qt6 interface, guided Home and separate Engine page
    * Safe updates, frozen per-job engine sources and asynchronous cancellation
    * Guided input preparation, verified imports and explicit project activation
    * Exact census IDs, validity masks, spatial-block validation and zone conservation
    * Preserved run history, quality diagnostics, styled QGIS maps and area summaries
    * Illustrated guides, deterministic packages and Windows/Linux regression tests
Version 0.1.1:
    * Added logarithmic scale option
    * Added configurable feature selection threshold
    * Improved configuration management
    * Fixed bugs in prediction processing
    * Enhanced UI layout
Version 0.1.0:
    * Initial release
    * Basic population modeling functionality
    * Support for parallel processing
    * Integration with QGIS layer system
tags=population, mapping, dasymetric, machine learning, GIS, Python
homepage=https://wpgp.github.io/QGIS-pypopRF
category=Raster
icon=icon.png
experimental=False
deprecated=False
server=False
