What lives here

auxiliary/ is a toolbox, not a service. Each subdirectory is an independent script (or pair of scripts) you run by hand on a workstation or build machine, before the data ever reaches MMGIS at runtime. The job of these tools is to take raw mission products — large GeoTIFFs in non-Earth projections, DEMs, GeoJSON dumps, colormap definitions — and turn them into the tiled, indexed, web-friendly artifacts MMGIS expects to find on disk or in Postgres.

Almost everything is Python wrapping GDAL/OGR. The handful of Node scripts (populateMosaics, geojson2ndgeojson, compileGeologic) cover the cases where streaming JSON or asset compilation is more natural than a GDAL pipeline. There is also a leftover csh helper (gis_parm/) for environment plumbing.

Rough groups

The scripts cluster by purpose:

• Raster to tiles for arbitrary projections. gdal2customtiles/ and gdal2tiles4extent/ are forks of the upstream gdal2tiles.py that add an --extentworld flag so you can slice rasters in IAU-2000 planetary projections (Lunar South Polar, Mars equirectangular, etc.) instead of being limited to Web Mercator. rasterstotiles/ and 1bto4b/ are convenience wrappers and bit-depth-conversion variants.
• DEM tiles. demtiles/ and the --dem mode of gdal2customtiles.py encode 32-bit elevation values into RGBA PNG tiles that the frontend's 3D globe and Viewshed tool decode at draw time. The encoding is a fixed convention — see the gdal2customtiles readme for the 0-value remapping detail.
• Bulk pipelines. bulk_tiles/ walks a directory of TIFFs, applies colormaps, generates tiles and legends in one shot, and can spit out JSON snippets ready to paste into an MMGIS layer config.
• Mosaics and STAC. populateMosaics/ decorates a GeoJSON of rover positions with image-tile URLs derived from a CSV of mosaic parameters. stac/tifs2cogs/ and stac/create-stac-items/ produce Cloud-Optimized GeoTIFFs and STAC items consumed by the adjacent TiTiler/STAC services.
• Vector normalization. geojson2ndgeojson/ rewrites a FeatureCollection as newline-delimited features so that very large vector dumps can be stream-uploaded into the Datasets module instead of buffered whole.
• Legends and colormaps. rastertolegend/ reads a GDAL color-relief file and emits both a colored raster and a legend.csv MMGIS can render. quantize_colormap/ derives a discrete matplotlib colormap from raster value distributions. rasterLabels/ rasterizes a labeled vector layer into a lookup raster.
• Domain-specific assets. compileGeologic/ is a one-off Node tool that bakes the FGDC geologic pattern/symbol library into PNGs and a JSON manifest the Draw tool consumes.

How they fit into mission-data prep

A typical onboarding flow for a new mission basemap is: drop a planetary-projected GeoTIFF into a working directory, run gdal2customtiles.py -p raster --extentworld … to produce a tile pyramid under Missions/<name>/Layers/, optionally run rastertolegend.py to produce a legend CSV alongside it, and then point an MMGIS layer config at the resulting directory. For vector data the path is similar: convert to ndGeoJSON, then POST it to the Datasets endpoints described in feature modules. Mosaic and STAC products feed the corresponding backend modules and the adjacent Python services.

Runtime environment

These scripts share the project's mmgis micromamba environment, defined at the repo root in python-environment.yml and built into the Docker image (see the WITH_STAC block in Dockerfile). It pins GDAL 3.9, rasterio, and the Python deps the STAC/TiTiler sidecars also depend on, so a single micromamba activate mmgis covers both auxiliary script runs and adjacent-server development. Per-script READMEs note any extra Node packages (csv-parser, JSONStream) that need an npm install first.