[general]
name=geobr
qgisMinimumVersion=3.40
qgisMaximumVersion=4.99
description=Download official spatial data sets of Brazil as QGIS layers
about=Easy access to official spatial data sets of Brazil. The plugin offers a wide range of official spatial data sets available at various geographic scales and for various years with harmonized attributes, projection and fixed topology. This includes states, municipalities, census tracts, biomes, indigenous lands, health facilities, schools, favelas and more. All original data sets come from IBGE and other official Brazilian sources and are returned in EPSG:4674 (SIRGAS 2000). This QGIS plugin is built on top of the geobr Python library, which integrates seamlessly with DuckDB and geopandas, and can be used from the Processing Toolbox, chained in Model Builder, run in batch, or called from qgis_process. The geobr library is bundled with the plugin, so there is nothing to pip-install. DuckDB, the engine geobr runs on, is installed for you by the qpip plugin, which QGIS offers to install alongside this one; if you skip qpip, install it into the Python QGIS uses: python -m pip install --user "duckdb>=1.5.3". Everything else geobr needs ships with the official QGIS builds for Windows and macOS; a Linux distribution build of QGIS may lack geopandas or pyarrow, and the plugin says so at startup.
version=0.4.1
author=Rafael H. M. Pereira, Ipea
email=rafa.pereira.br@gmail.com
repository=https://github.com/ipea/geobr
tracker=https://github.com/ipea/geobr/issues
homepage=https://github.com/ipea/geobr/tree/master/qgis-plugin
category=Vector
icon=icon.png
experimental=False
deprecated=False
hasProcessingProvider=True
plugin_dependencies=qpip
external_deps=duckdb>=1.5.3
tags=brazil,ibge,ipea,boundaries,census,municipality,opendata,official data,vector
changelog=0.4.1 - Security scan fixes in the bundled geobr
    The QGIS plugin repository's Bandit scan blocked 0.4.0 over 16 findings
    in the bundled geobr library. This release bundles geobr 2.1.1, which
    resolves all of them: the one genuine issue (data downloads no longer
    skip TLS certificate verification), a missing request timeout, and the
    string-built SQL and silent except blocks the scanner flags, rewritten
    with DuckDB's relational API and debug logging. No change in the data,
    the algorithms, or their parameters.
    0.4.0 - geobr bundled, duckdb through qpip, nothing to pip-install
    The geobr Python library (2.1.0) now ships inside the plugin, under
    _vendor/, and the plugin puts it on the Python path when it loads. The
    manual "pip install geobr" step of earlier releases is gone: install the
    plugin, accept qpip when the Plugin Manager offers it, done.
    duckdb, the engine geobr runs on, arrives through the qpip plugin: the
    plugin declares qpip as a dependency and ships a requirements.txt asking
    for duckdb 1.5.3 or newer. Installing geobr from the Plugin Manager
    offers to install qpip as well, and qpip then installs duckdb into your
    QGIS profile the first time the plugin loads. QDuckDB uses the same
    mechanism, so the two plugins share one duckdb.
    qpip only runs inside the QGIS desktop app, so the plugin adds qpip's
    dependency folder to the Python path itself; a qpip-installed duckdb
    therefore also works from qgis_process.
    The bundled copy takes precedence over a geobr that was pip-installed
    into QGIS's Python by an earlier version of these instructions, so the
    plugin always runs the geobr it was tested with. Upgrading geobr means
    upgrading the plugin. After upgrading the plugin, restart QGIS: Python
    keeps the previously imported library for the rest of the session.
    The startup warning and the run-time error say which piece is missing
    and where it comes from: the bundle, duckdb via qpip, or - on a Linux
    distribution build of QGIS that does not ship geopandas or pyarrow - the
    package to install for the Python QGIS uses.
    geobr 2.1.0 brings the new Addresses algorithm (CNEFE 2022 geolocated
    addresses, per municipality) and raises an error on an invalid code
    filter instead of returning the whole country; the plugin still checks
    codes up front, which spares the download.
    0.3.1 - Layer names
    Layers are now named after the data they hold - municipalities_2020,
    healthfacilities_202504, censustracts_2000_urban,
    healthregions_2013_macro - instead of every algorithm producing an
    identically named layer that collided with the last one. A destination
    filename you choose yourself still names the layer, as before.
    The "Simplified" checkbox now reads "Simplified geometry", which is what
    it actually selects.
    The year and date fields are now a drop-down of the years each data set
    offers, newest first and preselected, read from geobr's own metadata
    when the dialog is first opened, instead of an empty optional number.
    The algorithm window opens at least 700 px wide, instead of the
    Processing default that squeezed the parameters beside the help panel.
    It can still be resized, and the size is remembered as usual.
    Requires geobr 2.0.1 or newer. Earlier releases pinned geopandas and
    shapely below the versions QGIS ships, so installing geobr downgraded
    QGIS's own copies of both.
    0.3.0 - First stable release
    Same code as 0.2.1, released without the experimental flag so the
    plugin appears in the QGIS Plugin Manager without "Show also
    experimental plugins" enabled.
    0.2.1 - Qt6 and an empty year field
    The year and date fields now open empty instead of on 1872; a missing
    value is reported before anything is downloaded.
    Enum access follows the Qt6 spelling flagged by the plugin repository's
    Qt6 compatibility check (QGIS 4).
    0.2.0 - Built on geobr 2.0.0
    Requires geobr 2.0.0 or newer.
    Restored the algorithm help text, which geobr 2.0.0 broke by moving its
    shared parameter documentation into an import-time decorator.
    Health regions now offer the geometry_level control (municipality, micro,
    macro), which geobr 2.0.0 fixed.
    Reader arguments follow geobr 2.0.0's order, which now matches the R
    package, so parameters appear in a different order than in 0.1.0.
    Rejects code filters geobr would silently ignore, and mixed code lists it
    would only partly apply.
    0.1.0 - First release
    Exposes geobr, a computational package to download official spatial data sets
    of Brazil. The package covers a wide range of spatial data sets,
    available at various geographic scales and for various years with
    harmonized attributes, projection and fixed topology. This QGIS plugin 
    exposes the geobr Python package as processing algorithms. Every geobr 
    reader — states, municipalities, census tracts, biomes, indigenous lands, 
    health facilities, schools, favelas, polling places and more — becomes 
    an algorithm in the Processing Toolbox.
