KARTOZA · INFRASTRUCTURE MAPPER

A versioned spatial schema for site infrastructure

Capture in Postgres, snapshot to GeoPackage, sync from the field — one source of truth.

Get Started Data model GitHub

What it is

Infrastructure Mapper is an opinionated PostgreSQL/PostGIS data model for capturing buildings, roads, water, electricity, fences, vegetation, monitoring, points of interest, land use, gates, poles, and culinary facilities — together with every condition, material, and reading observed against them over time.

Ship the schema as a Postgres database. Snapshot it as a GeoPackage. Take it to the field in QField or Mergin Maps. Migrate it forward with vX.Y.Z migrations. All from the same source of truth.

Download the latest release

Always-current direct links — no version pinning, no GitHub navigation:

• GeoPackage

  ---

  Single .gpkg with every domain. Drag onto a QGIS canvas.

  KartozaInfrastructureMapper-latest.gpkg

• PostgreSQL schema

  ---

  Self-contained SQL: extensions + meta + every domain + fixtures.

  pg-schema-latest.sql

• Migrations

  ---

  Forward migrations + runner scripts. Apply on top of an older DB.

  pg-migrations-latest.tar.gz · gpkg-migrations-latest.tar.gz

Want just one domain? Every release also ships 13 per-domain SQL + GeoPackage slices (e.g. pg-schema-07-fencing-latest.sql, KartozaInfrastructureMapper-07-fencing-latest.gpkg) — see Getting Started for the per-domain download patterns, or GitHub Releases for the full asset list.

What's in the box

• Every capture domain

  ---

  From sub-soil readings to gate hinge angles — everything you need to map a working site, modelled relationally with stable UUIDs across PG and GPKG.

  Explore the data model

• Versioned schema lifecycle

  ---

  Immutable baseline plus paired PG and GPKG migration trails. Strict-sequential runners apply pending versions in semver order, each in its own transaction.

  Read the lifecycle

• Field-ready GeoPackages

  ---

  The same schema, snapshot to a single file for QField or Mergin Maps. UUIDs and last_update timestamps make a clean sync path back to the source PG.

  Field workflow

• One-tag releases

  ---

  Each vX.Y.Z tag publishes a fresh PG schema dump, GPKG, fixtures, and migration tarballs as GitHub Release artifacts.

  Release process

• Reproducible Nix dev shell

  ---

  One flake provisions Postgres, PostGIS, QGIS, GDAL, sqlfluff, black, markdownlint, mkdocs — everything CI uses, ready locally with nix develop.

  The Nix flake

• CI that knows the schema

  ---

  Pre-commit + GitHub Action enforce baseline immutability and the -- Issue-NNN: annotation convention. The schema cannot drift silently.

  CI & pre-commit

QA status

Made with 💗 by Kartoza · Sponsor on GitHub · Repository