Awesome

Static Map Renderer using Mapbox GL

Create static map images using MapLibre GL with a command-line interface, an HTTP interface, and a NodeJS API.

Features:

• Render static maps using NodeJS with maplibre-gl-native
• Supports raster and vector tiles
• Compatible with Mapbox tiles (don't forget attribution) and other hosted tile providers
• Use locally hosted mbtiles
• Add GeoJSON overlays to your maps
• Supports high DPI rendering
• Also available for use in Docker

Blog post describing the background and goals in a bit more detail.

One of the nifty features of this package is that you can use locally hosted mbtiles files with your raster or vector tiles. This saves considerable time during rendering compared to using map services over the web.

This package is intended to help with generating static maps for download or use in reports, especially when combined with your own styles or overlays.

If you are only using hosted Mapbox styles and vector tiles, please use the Mapbox Static API instead; it is more full featured and more appropriate for static Mapbox maps.

Attribution

Please make sure to give appropriate attribution to the data sources and styles used in your maps, in the manner that those providers specify.

If you use Mapbox styles or hosted tiles, make sure to include appropriate attribution in your output maps.

Installation

npm add mbgl-renderer or npm install mbgl-renderer

Supported versions of NodeJS:

• 16
• 18

Only NodeJS versions with @maplibre/maplibre-gl-native binaries built by MapLibre are supported via npm install, otherwise you need to build @maplibre/maplibre-gl-native from source yourself. See build instructions for more information.

On a server, in addition to build tools, you need to install a GL environment. See the Dockerfile and entrypoint.sh for an example setup.

Usage

Style JSON:

The primary input to every map rendering method below is a Mapbox GL Style JSON file. For full reference on this format, please see the Mapbox documentation.

Labels

In order to use text labels, you need to include glyphs in your Style JSON (sibling of sources)

To use Mapbox hosted glyphs (a Mapbox token is required):

"glyphs": "mapbox://fonts/mapbox/{fontstack}/{range}.pbf"

You can also use OpenMapTiles hosted glyphs:

"glyphs": "http://fonts.openmaptiles.org/{fontstack}/{range}.pbf",

In either case, you must make sure that the font you specify is available from that provider. See tests/fixtures/example-style-geojson-label.json for an example of adding labels based on coordinates specified in GeoJSON.

NodeJS API:

The following examples assume that you are using babel to provide ES6 features.

import render from 'mbgl-renderer'

import style from `tests/fixtures/example-style.json`
// style JSON file with MapBox style.  Can also be opened and read instead of imported.

const width = 512
const height = 256
const center = [-79.86, 32.68]
const zoom = 10

render(style, width, height, { zoom, center })
    .then((data) => {
        fs.writeFileSync('test.png', data)
    }))

You can also provide bounds instead of center and zoom:

const width = 512
const height = 256
const bounds = [-80.23, 32.678, -79.73, 32.891]

render(style, width, height, { bounds })
    .then((data) => {
        fs.writeFileSync('test.png', data)
    }))

If you provide bounds you can also provide padding to add that many pixels to each side of the rendered image. These pixels are padded to the inside of the image, meaning that the resulting image matches the width and height you provide, but is zoomed out to allow padding around the edges.

const width = 512
const height = 256
const bounds = [-80.23, 32.678, -79.73, 32.891]
const padding = 25

render(style, width, height, { bounds, padding })
    .then((data) => {
        fs.writeFileSync('test.png', data)
    }))

padding must be integers and not be greater than 1/2 of width or height, whichever is smaller. You can provide a negative padding to over-zoom the image.

You can also supply a pixel ratio for High DPI screens, typically > 1 (max of 31 has been tested):

const width = 512
const height = 256
const center = [-79.86, 32.68]
const zoom = 10
const ratio = 2

render(style, width, height, { zoom, center, ratio })
    .then((data) => {
        fs.writeFileSync('test.png', data)
    }))

You can also provide an alternative bearing (0-360) or pitch (0-60):

const width = 512
const height = 256
const center = [-79.86, 32.68]
const zoom = 10
const bearing = 90
const pitch = 30

render(style, width, height, { zoom, center, bearing, pitch })
    .then((data) => {
        fs.writeFileSync('test.png', data)
    }))

If your style includes a Mapbox hosted source (e.g., "url": "mapbox://mapbox.mapbox-streets-v7"), you need to pass in your Mapbox access token as well:

render(style, width, height, { bounds, token: '<your access token>' })
    .then((data) => {
        fs.writeFileSync('test.png', data)
    }))

You can also provide icon images that will be available for icon-image, line-pattern, and fill-pattern style properties. The images property is an object with keys for at least url and optional keys pixelRatio and sdf.

url must point to a valid remote URL that is accessible from mbgl-renderer. It may include base64 encoded image data (example below).

Set sdf to true to enable the image to be converted into an SDF icon. Read more.

pixelRatio is only used for symbols; due to a bug in Mapbox GL Native, it does not work properly for line-pattern and fill-pattern icon images.

const images = {
    "exampleImageID": {
        "url": "data:image/png;base64,iVBORw0KGg....truncated...",
        "sdf": true / false (default),
        "pixelRatio": 1 (default)
    }
}

Then refer to exampleImageID in your style:

...
layers:[
    {
        "id": "...",
        "type": "symbol",
        "source": "...",
        "layout": {
            "icon-image": "exampleImageID",
            ...
        },
        "paint": {
            ...
        }
    }
]

Command line interface:

  Usage: mbgl-render <style.json> <img_filename> <width> <height> [options]

  Export a Mapbox GL map to image.  You must provide either center and zoom, or bounds.

  Options:

    -V, --version                         output the version number
    -c, --center <longitude,latitude>     center of map (NO SPACES)
    -z, --zoom <n>                        Zoom level
    -r, --ratio <n>                       Pixel ratio
    -b, --bounds <west,south,east,north>  Bounds (NO SPACES)
    --padding <padding>                   Number of pixels to add to the inside of each edge of the image.  Can only be used with bounds option.
    --bearing <degrees>                   Bearing in degrees (0-360)
    --pitch <degrees>                     Pitch in degrees (0-60)
    -t, --tiles <mbtiles_path>            Directory containing local mbtiles files to render
    --token <mapbox access token>         Mapbox access token (required for using Mapbox styles and sources)
    --images <images.json                 JSON file containing image config
    -h, --help                            output usage information

To render an image using center (NO spaces or brackets) and zoom:

mbgl-render tests/fixtures/example-style.json test.png 512 256 -c -79.86,32.68 -z 10

To render an image using bounds (NO spaces or brackets):

mbgl-render tests/fixtures/example-style.json test.png 512 256 -b -80.23,32.678,-79.73,32.891

To render an image using bounds and padding:

mbgl-render tests/fixtures/example-style.json test.png 512 256 -b -80.23,32.678,-79.73,32.891 --padding 25

To use local mbtiles tilesets:

mbgl-render tests/fixtures/example-style-mbtiles-source-vector.json test.png 1024 1024 -z 0 -c 0,0 -t tests/fixtures

To use a Mapbox hosted style (see attribution above!):

mbgl-render mapbox://styles/mapbox/outdoors-v10 test.png 1024 1024 -c 0,0 -z 0 --token <your mapbox token>

Note: support for Mapbox hosted styles is still considered experimental.

Static image server

You start this from the command line:

  Usage: mbgl-server [options]

  Start a server to render Mapbox GL map requests to images.

  Options:

    -V, --version               output the version number
    -p, --port <n>              Server port
    -t, --tiles <mbtiles_path>  Directory containing local mbtiles files to render
    -v, --verbose               Enable request logging
    -h, --help                  output usage information

To start this on port 8080 with local tiles in tests/fixtures:

mbgl-static-server -p 8080 -t tests/fixtures

You can also start this via npm start but you must use the -- spacer before passing argmuents:

npm start -- --port 8080 --tiles tests/fixtures

Making requests

In your client of choice, you can make either HTTP GET or POST requests to

http://localhost:8080/render

GET requests:

Include the following query parameters:

• height and width are integer values (required).
• zoom is a floating point value.
• ratio is an integer value.
• center if provided must be a longitude,latitude with floating point values (NO spaces or brackets).
• bounds if provided must be west,south,east,north with floating point values (NO spaces or brackets).
• padding if provided must be an integer value that is less than 1/2 of width or height, whichever is smaller. Can only be used with bounds.
• bearing is a floating point value (0-360)pitch`is a floating point value (0-60).
• token if provided must a string.
• images if provided must be URL encoded JSON object (see images property above).

Your style JSON needs to be URL encoded:

http://localhost:8080/render?height=1024&width=1024&center=-80,-20&zoom=3&style=%7B%22version%22%3A8%2C%22sources%22%3A%7B%22land%22%3A%7B%22type%22%3A%22vector%22%2C%22url%22%3A%22mbtiles%3A%2F%2Fland%22%2C%22tileSize%22%3A256%7D%7D%2C%22layers%22%3A%5B%7B%22id%22%3A%22land%22%2C%22type%22%3A%22fill%22%2C%22source%22%3A%22land%22%2C%22source-layer%22%3A%22land%22%2C%22paint%22%3A%7B%22fill-color%22%3A%22%23AAAAAA%22%2C%22fill-opacity%22%3A1%7D%7D%5D%7D

If your style JSON points to local tilesets, you must have started the server up using those local tilesets.

To test the server from the command line, for example with the excellent tool httpie with the style file tests/fixtures/example-style-mbtiles-source.json:

http :8080/render width=400 height=400 zoom=0 center=0,0 style:=@tests/fixtures/example-style.json > /tmp/test.png

POST requests:

You can do a POST request with all of the above parameters in the body of the request, and the style can be passed directly as JSON instead of URL encoded.

POST may be necessary where your style JSON file exceeds the maximum number of characters allowed in a GET request URL.

Development

Use npm run watch to start up a file watcher to recompile ES6 files in src/ to ES5 files that are executable in Node in dist/. These are compiled using babel.

Testing

Tests are run using jest. Right now, our coverage is not great and tests only exercise the core functionality of the render function.

Tests require a valid Mapbox API token. Set this in .env.test file in the root of the repository:

MAPBOX_API_TOKEN=<your token>

To run tests:

npm run test

This uses the pixelmatch package to determine if output images match those that are expected. This may fail when rendered on different machines for reasons we have not completely sorted out, so don't necessarily be alarmed that tests are failing for you - check the outputs.

Docker

Pull the latest image from Github Container Registry:

docker pull ghcr.io/consbio/mbgl-renderer:latest

To run mbgl-server in the docker container on port 8080:

docker run --rm -p 8080:80 consbio/mbgl-renderer

Mount your local tiles directory to /app/tiles in the container to use these in the server or CLI:

docker run --rm -p 8080:80 -v $(pwd)/tests/fixtures:/app/tiles consbio/mbgl-renderer

Build your own image:

Build your own docker container with name <image_name>:

docker build -t <image_name> -f docker/Dockerfile .

Headless servers

In order to use this package on a headless server, you need to use xvfb. See docker/Dockerfile and docker/entrypoint.sh for the basic configuration.

Credits

This project was made possible based on support from the South Atlantic Landscape Conservation Cooperative, the Paulson Institute, and the California Department of Food and Agriculture - Office of Environmental Farming & Innovation.

Contributors ✨

Thanks goes to these wonderful people (emoji key):

This project follows the all-contributors specification. Contributions of any kind welcome!