mlt_py

Python bindings for the MapLibre Tile (MLT) format via PyO3.

Decoding

import maplibre_tiles

data = open("tile.mlt", "rb").read()

# Structured decode with geo-referencing
layers = maplibre_tiles.decode_mlt(data, z=14, x=8297, y=10749, tms=True)
for layer in layers:
    print(f"Layer: {layer.name}, extent: {layer.extent}")
    for feat in layer.features[:3]:
        print(f"  id={feat.id}, type={feat.geometry_type}")
        print(f"  wkb={len(feat.wkb)} bytes, props={dict(feat.properties)}")

# Raw tile-local coordinates (no z/x/y needed)
layers = maplibre_tiles.decode_mlt(data)

# GeoJSON string output (tile-local coords)
geojson_str = maplibre_tiles.decode_mlt_to_geojson(data)

# Fast layer listing (no full decode)
names = maplibre_tiles.list_layers(data)

Decoded objects

decode_mlt(...) returns a list of MltLayer objects. Each MltLayer represents one decoded MLT layer and exposes:

• name: str — the layer name.
• extent: int — the layer extent.
• features: list[MltFeature] — the decoded features in that layer.

Each MltFeature represents one decoded feature and exposes:

• id: int | None — the feature id, if present.
• geometry_type: str — the decoded geometry type.
• wkb: bytes — the geometry as WKB.
• properties: dict — the feature properties.

import maplibre_tiles


layers = maplibre_tiles.decode_mlt(data)
for layer in layers:
    print(layer.name, layer.extent)
    for feature in layer.features[:2]:
        print(feature.id, feature.geometry_type, dict(feature.properties))

Encoding

encode_geojson(geojson, name, extent=4096, *, tessellate=False, sort="auto", shared_dict=True, fsst=True, fastpfor=True) -> bytes encodes a single layer to an MLT blob.

• geojson is a GeoJSON FeatureCollection. Geometry is in tile-local coordinate space (no projection), matching tilezen/mapbox-vector-tile's default. Coordinates must be integers and 2D. They must be JSON integers (2048), not floats: a float-typed value such as 2048.0 raises ValueError.
• name and extent set the MLT layer metadata, since a FeatureCollection has no slot for them. extent defaults to 4096.
• tessellate generates triangulation data for polygons and multi-polygons.
• sort controls which feature ordering(s) the encoder tries: all, auto, morton, hilbert, id, or none.
• shared_dict enables grouping strings into shared dictionaries.
• fsst enables FSST string compression.
• fastpfor enables FastPFOR integer compression.

encode_mvt(data, *, tessellate=False, sort="auto", shared_dict=True, fsst=True, fastpfor=True) -> bytes encodes an entire raw Mapbox Vector Tile (protobuf) to MLT using the same encoding options.

import maplibre_tiles


blob = maplibre_tiles.encode_geojson(
    {
        "type": "FeatureCollection",
        "features": [
            {
                "type": "Feature",
                "id": 1,
                "geometry": {"type": "Point", "coordinates": },
                "properties": {"name": "main", "lanes": 3},
            },
        ],
    },
    name="roads",
    extent=4096,
)


# Multi-layer tiles -> encode each layer and concatenate the bytes
tile = b"".join([
    maplibre_tiles.encode_geojson(roads, name="roads"),
    maplibre_tiles.encode_geojson(water, name="water"),
])

import maplibre_tiles


mvt = open("tile.mvt", "rb").read()

blob = maplibre_tiles.encode_mvt(mvt)


# With explicit encoding options
blob = maplibre_tiles.encode_mvt(
    mvt,
    tessellate=False,
    sort="auto",
    shared_dict=True,
    fsst=True,
    fastpfor=True,
)

Input is validated strictly. A non-FeatureCollection input, a non-Feature member, float-typed or 3D coordinates, null or empty geometry, nested or non-scalar property values, a non-u64 id, and an empty collection all raise ValueError.