Skip to content

Open Topo Data Server Documentation

Getting started

The easiest way to run Open Topo Data is with Docker. Install docker then run the following commands:

git clone https://github.com/ajnisbet/opentopodata.git
cd opentopodata
make build
make run

This will start a server on localhost:5000 with a small demo dataset called test-dataset. Check out the API docs for info about the format of requests and responses.

Getting started on M1 / Apple Silicon Macs

On M1 Macs you'll probably need to use the alternate apple-silicon.Dockerfile docker image, which includes build dependencies for libraries that don't have binaries for M1.

git clone https://github.com/ajnisbet/opentopodata.git
cd opentopodata
make build-m1
make run

This should work without rosetta. See issue #55 for more info.

Getting started on Windows

On Windows you'll probably need to run the build and run commands without make:

git clone https://github.com/ajnisbet/opentopodata.git
cd opentopodata
docker build --tag opentopodata --file docker/Dockerfile . 
docker run --rm -it --volume C:/path/to/opentopodata/data:/app/data:ro -p 5000:5000 opentopodata sh -c "/usr/bin/supervisord -c /app/docker/supervisord.conf"

For more details see this note on windows support.

Dataset support

Open Topo Data supports all georeferenced raster formats supported by GDAL (e.g, .tiff, .hgt, .jp2).

Datasets can take one of two formats:

  • A single raster file.
  • A collection of square raster tiles which follow the SRTM naming convention: the file is named for the lower left corner. So a file named N30W120.tiff would span from 30 to 31 degrees latitude, and -120 to -119 degrees longitude. By default tiles are 1° by 1° and the coordinates are in WGS84, but this can be configured.

If your dataset consists of multiple files that aren't on a nice grid, you can create a .vrt file pointing to the files that Open Topo Data will treat as a single-file dataset. For an example of this process, see the documentation for configuring EMODnet.

Configuration

Open Topo Data is configured by a config.yaml file. If that file is missing it will look for example-config.yaml instead. You can set the CONFIG_PATH environment variable to specify a different path.

A config might look like:

max_locations_per_request: 100 
access_control_allow_origin: '*'
datasets:
- name: etopo1
  path: data/etopo1/
- name: srtm90m
  path: data/srtm-90m-v3/
  filename_epsg: 4326
  filename_tile_size: 1

corresponding to a directory structure:

opentopodata
|
└───data
    |
    ├───etopo1
    |   |
    |   └───etopo1-dem.geotiff
    |
    └───srtm-90m-v3
        |
        ├───N00E000.hgt 
        ├───N00E001.hgt 
        ├───N00E002.hgt 
        ├───etc...

which would expose localhost:5000/v1/etopo1 and localhost:5000/v1/srtm90m.

Mofifying the config file triggers a restart of OpenTopoData (which will reload the new config).

Config spec

  • max_locations_per_request: Requests with more than this many locations will return a 400 error. Default: 100.
  • access_control_allow_origin: Value for the Access-Control-Allow-Origin CORS header. Set to * or a domain to allow in-browser requests from a different origin. Set to null to send no Access-Control-Allow-Origin header. Default: null.
  • datasets[].name: Dataset name, used in url. Required.
  • datasets[].path: Path to folder containing the dataset. If the dataset is a single file it must be placed inside a folder. This path is relative to the repository directory inside docker. I suggest placing datasets inside the provided data folder, which is mounted in docker by make run. Files can be nested arbitrarily inside the dataset path. Required.
  • datasets[].filename_epsg: For tiled datasets, the projection of the filename coordinates. The default value is 4326, which is latitude/longitude with the WGS84 datum.
  • datasets[].filename_tile_size: For tiled datasets, how large each square tile is in the units of filename_epsg. For example, a lat,lon location of 38.2,121.2 would lie in the tile N38W121 for a tile size of 1, but lie in N35W120 for a tile size of 5. For non-integer tile sizes like 2.5, specify them as a string to avoid floating point parsing issues: "2.5". Default: 1.
  • datasets[].wgs84_bounds.left: Leftmost (westmost) longitude of the dataset, in WGS84. Used as a performance optimisation for Multi datasets. Default: -180.
  • datasets[].wgs84_bounds.right: Rightmost (eastmost) longitude of the dataset. Default: 180.
  • datasets[].wgs84_bounds.bottom: Bottommost (southmost) latitude of the dataset. Default: -90.
  • datasets[].wgs84_bounds.top: Topmost (northmost) latitude of the dataset. Default: 90.
  • datasets[].child_datasets[]: A list of names of other datasets. Querying this MultiDataset will check each dataset in child_datasets in order until a non-null elevation is found. For more information see Multi datasets.

Adding datasets

An important goal of Open Topo Data is make is easy to add new datasets. The included dataset is very low resolution (etopo1 downsampled to about 100km) and is intended only for testing.

Adding a new dataset takes two steps:

  1. placing the dataset in the data directory
  2. adding the path to the dataset in config.yaml.

Instructions are provided for adding the various datasets used in the public API:

Kubernetes

See How to deploy on Kubernetes for details and config files for running on kubernetes.