STAC Catalog

STAC Catalog (stac-catalog) is a Python toolkit that turns structured raster metadata — a CSV file or a Google Sheet — into valid Spatio-Temporal Asset Catalog (STAC) Collections and Items, loads them into pgSTAC, and serves them through a ready-made Docker stack (pgSTAC + STAC FastAPI + TiTiler + STAC Browser).

It is not a browser or a fork of stac-browser. It is the ingestion pipeline that builds a clean, STAC-compliant catalog from the metadata you already keep in spreadsheets, so you can browse and query your rasters without hand-writing STAC JSON.

:::info Where STAC Browser fits The Docker stack bundles the upstream radiantearth/stac-browser as the human-facing front end, plus TiTiler for raster rendering and previews. stac-catalog is the tool that generates and loads the data those services serve. :::

How it works

CSV ────────────┐
                ├──> normalized metadata (DataFrames)
Google Sheets ──┘                ↓
                            stac-catalog
                                 ↓
                     Collections + Items (STAC JSON)
                                 ↓
                               pgSTAC
                                 ↓
                         STAC FastAPI
                        ↙        ↘
                  STAC Browser   TiTiler

CSV and Google Sheets are just source adapters: both are normalized into the same internal tables before generation, so the STAC builder does not depend on where your metadata lives.

Features

• Two metadata sources — CSV files and Google Sheets, normalized to the same internal model.
• Collections & Items — grouped and non-grouped datasets, including grouped Data Cube variables and dimensions.
• COG inspection — reads Cloud Optimized GeoTIFFs to derive geometry, CRS, resolution, dtype, nodata, scale, offset, file size, and checksum.
• Filename-based parsing — derives product, statistic, resolution, dates, region, EPSG, version, depth, and variants from conventional raster names.
• Data Cube metadata — variables, depth (z) axis, positional statistics, and variants.
• Rich STAC fields — DOI, citation, providers, catalogs, and more.
• Style assets — QML and SLD linked to matching raster variables.
• TiTiler integration — thumbnails, colormaps, and rescaling written into the STAC assets.
• Validation — input metadata is validated before generation, with clear row-level error messages.
• Loading — pgSTAC upsert via stac-catalog load, or a generate-and-load one-step stac-catalog sync.
• Production loading — run database-dependent commands inside the private Docker network with --prod, keeping PostgreSQL off the public network.
• Docker Compose stack — pgSTAC, STAC FastAPI, TiTiler, and STAC Browser in one docker compose up.

How it fits into STAC GIS

Scenario	Setup
Publish a curated catalog	Point your rasters' metadata at stac-catalog, generate + load, and serve through the bundled STAC stack.
Feed STAC GIS Core	Load the generated Collections/Items into Core's pgSTAC/STAC API, or run the stack standalone.
Self-host a public catalog	Expose STAC FastAPI + STAC Browser + TiTiler behind a reverse proxy; keep the database private (see Deployment).

💡 STAC GIS Core is the full platform (storage, app, AI). The Catalog is the lightweight, spreadsheet-driven tool that produces STAC metadata and a serving stack. You can use the Catalog on its own to publish any set of rasters as a STAC catalog.

Documentation map

Page	What it covers
Installation	Install the stac-catalog package and its optional extras
Getting started	Local end-to-end run: start the stack, generate, load, browse
Configuration	Config file, precedence, sources, environment variables
CLI reference	generate, load, sync, every option, and --prod
Metadata	Input tables, filename convention, fields, grouping, dates, depth, styles
Deployment	The Docker stack, local vs. production, networking, persistence

License

STAC Catalog is licensed under the Apache License, Version 2.0.