mappyfile-gdal

A mappyfile plugin to create Mapfiles for GDAL datasets.

It takes the JSON output of gdal raster info or gdal vector info and writes a MapServer Mapfile for it, with a layer per dataset, a map extent and size matching the data, and default symbology so the result can be rendered straight away.

Installation

pip install mappyfile-gdal

GDAL 3.11 or later is needed for the gdal command line tool, and MapServer to render the generated Mapfiles. Both are available from conda-forge:

conda create -n gdal-mapserver -c conda-forge gdal mapserver --yes
conda activate gdal-mapserver
pip install mappyfile-gdal

Usage

Pipe gdal info output straight into the tool, using - to read from stdin:

gdal raster info --stats raster.tif | mappyfile-gdal - raster.map
map2img -m raster.map -o raster.png

Without an output path the Mapfile is printed to stdout:

gdal vector info poly.gpkg | mappyfile-gdal

Or use a saved JSON file, or a whole folder of them:

gdal raster info --stats raster.tif --of json > raster.json
mappyfile-gdal raster.json raster.map

mappyfile-gdal ./json ./mapfiles

Example

Two layers from the Natural Earth GeoPackage, piped through to a rendered image:

gdal info natural_earth_vector.gpkg --of JSON --layer ne_10m_lakes --layer ne_10m_admin_0_countries | mappyfile-gdal - out.map
map2img -m out.map -o out.png

The generated Mapfile is out.map.

Options

gdal info reports paths exactly as they were given to it, so a dataset opened with a relative path produces a Mapfile with a relative DATA. MapServer resolves those against SHAPEPATH, or against the Mapfile's own folder when SHAPEPATH isn't set:

gdal raster info --stats aaigrid/byte.asc | mappyfile-gdal - byte.map --shapepath /data/gdal

Statistics

Passing --stats to gdal raster info allows minimum and maximum to be written as PROCESSING "SCALE=min,max", which stretches the values to the full range of greys. Without it MapServer scales each request from the pixels it happens to be drawing, so the same value appears as different shades at different zoom levels.

Selecting layers

For a dataset with many layers, such as a GeoPackage, every layer is added to the Mapfile. To get a Mapfile for a single layer, select it in GDAL using the --layer option:

gdal vector info --layer roads data.gpkg | mappyfile-gdal - roads.map

Python API

import json

from mappyfile_gdal.main import main

with open("example.json", encoding="utf-8") as f:
    data = json.load(f)

# save the Mapfile and return its path
main(data, "example.map", shapepath="/data/gdal")

# or return the Mapfile as a string
mapfile = main(data)

What gets generated

• Rasters become a TYPE RASTER layer, with PROCESSING "SCALE" and "NODATA" when statistics are available.
• Vectors use CONNECTIONTYPE OGR with one layer per geometry layer, polygons are drawn first and points last, each in its own colour with semi-transparent fills so overlapping layers stay visible.
• WMS capabilities documents use CONNECTIONTYPE WMS with one layer per subdataset.
• EXTENT is inset by half a pixel, since MapServer measures it between the centers of the corner pixels while GDAL reports the outer edges.
• PROJECTION is set when the dataset's CRS has an authority code. Datasets without one are drawn in their native coordinates.

Development

git clone https://github.com/geographika/mappyfile-gdal.git
cd mappyfile-gdal
pip install -e .
pip install -r requirements-dev.txt

pytest
flake8 .
mypy mappyfile_gdal

License

MIT