queries.yaml

all:
  - water
  - earth
  - places
  - landuse
  - roads
  - buildings
  - pois
  - boundaries
  - transit

sources:

  # NOTE: each source represents a query to the database
  # within each source, a list of templates can be used
  # the templates will be combined with a sql `union all` in between
  # therefore, it's necessary that the query results of each template
  # have the same types

  # NOTE: end_zoom is exclusive

  planet_osm_polygon:
    - template: planet_osm_polygon.jinja2
      # landuse and pois start at 4
      start_zoom: 4
  planet_osm_line:
    - template: planet_osm_line.jinja2
      # transit starts at 5
      start_zoom: 5
  planet_osm_point:
    # earth and places start at zoom 0
    - template: planet_osm_point.jinja2

  ne:
    ## places
    # ne_10m_populated_places
    - template: ne-places.jinja2

    ## boundaries

    # ne_110m_admin_0_boundary_lines_land
    - template: ne-boundaries-110m.jinja2
      end_zoom: 2

    # ne_50m_admin_0_boundary_lines_land
    # ne_50m_admin_1_states_provinces_lines
    - template: ne-boundaries-50m.jinja2
      start_zoom: 2
      end_zoom: 5

    # ne_10m_admin_0_boundary_lines_land
    # ne_10m_admin_0_boundary_lines_map_units
    # ne_10m_admin_1_states_provinces_lines
    - template: ne-boundaries-10m.jinja2
      start_zoom: 5
      end_zoom: 8

    ## earth

    # ne_110m_land
    - template: ne-earth-110m.jinja2
      end_zoom: 2

    # ne_50m_land
    - template: ne-earth-50m.jinja2
      start_zoom: 2
      end_zoom: 5

    # ne_10m_land
    - template: ne-earth-10m.jinja2
      start_zoom: 5
      end_zoom: 8

    ## landuse

    # ne_50m_urban_areas
    - template: ne-landuse-50m.jinja2
      start_zoom: 4
      end_zoom: 6

    # ne_10m_urban_areas
    - template: ne-landuse-10m.jinja2
      start_zoom: 6
      end_zoom: 10

    ## roads

    # ne_10m_roads
    - template: ne-roads.jinja2
      start_zoom: 5
      end_zoom: 8

    ## water

    # ne_110m_ocean
    # ne_110m_lakes
    # ne_110m_coastline
    # ne_110m_lakes
    - template: ne-water-110m.jinja2
      end_zoom: 2

    # ne_50m_ocean
    # ne_50m_lakes
    # ne_50m_coastline
    # ne_50m_lakes
    - template: ne-water-50m.jinja2
      start_zoom: 2
      end_zoom: 5

    # ne_10m_ocean
    # ne_10m_lakes
    # ne_10m_coastline
    # ne_10m_lakes
    - template: ne-water-10m.jinja2
      start_zoom: 5
      end_zoom: 8

    # ne_50m_playas
    - template: ne-water-playas-50m.jinja2
      start_zoom: 4
      end_zoom: 5

    # ne_10m_playas
    - template: ne-water-playas-10m.jinja2
      start_zoom: 5
      end_zoom: 8

  osmdata:

    - template: osmdata-earth.jinja2
      start_zoom: 8

    - template: osmdata-water.jinja2
      start_zoom: 8

  wof:
    - template: wof.jinja2
      start_zoom: 11

layers:
  water:
    template: water.jinja2
    geometry_types: [Point, MultiPoint, Polygon, MultiPolygon, LineString, MultiLineString]
    simplify_before_intersect: true
    simplify_start: 9
    transform:
      - vectordatasource.transform.tags_create_dict
      - vectordatasource.transform.tags_name_i18n
      - vectordatasource.transform.tags_remove
      - vectordatasource.transform.water_tunnel
      - vectordatasource.transform.add_id_to_properties
      - vectordatasource.transform.detect_osm_relation
      - vectordatasource.transform.remove_feature_id
      - vectordatasource.transform.truncate_min_zoom_to_2dp
    sort: vectordatasource.sort.water
    area-inclusion-threshold: 1
  earth:
    template: earth.jinja2
    geometry_types: [Point, MultiPoint, Polygon, MultiPolygon, LineString, MultiLineString]
    simplify_before_intersect: true
    simplify_start: 9
    transform:
      - vectordatasource.transform.tags_create_dict
      - vectordatasource.transform.tags_name_i18n
      - vectordatasource.transform.tags_remove
      - vectordatasource.transform.detect_osm_relation
      - vectordatasource.transform.add_id_to_properties
      - vectordatasource.transform.remove_feature_id
      - vectordatasource.transform.truncate_min_zoom_to_2dp
    sort: vectordatasource.sort.earth
    area-inclusion-threshold: 1
  places:
    template: places.jinja2
    geometry_types: [Point, MultiPoint]
    transform:
      - vectordatasource.transform.tags_create_dict
      - vectordatasource.transform.tags_name_i18n
      - vectordatasource.transform.tags_remove
      - vectordatasource.transform.place_population_int
      - vectordatasource.transform.calculate_default_place_min_zoom
      - vectordatasource.transform.add_id_to_properties
      - vectordatasource.transform.detect_osm_relation
      - vectordatasource.transform.remove_feature_id
      - vectordatasource.transform.truncate_min_zoom_to_2dp
    sort: vectordatasource.sort.places
    area-inclusion-threshold: 1
  landuse:
    template: landuse.jinja2
    geometry_types: [Polygon, MultiPolygon, LineString, MultiLineString]
    simplify_start: 4
    transform:
      - vectordatasource.transform.tags_create_dict
      - vectordatasource.transform.tags_name_i18n
      - vectordatasource.transform.tags_remove
      - vectordatasource.transform.normalize_tourism_kind
      - vectordatasource.transform.add_id_to_properties
      - vectordatasource.transform.detect_osm_relation
      - vectordatasource.transform.remove_feature_id
      - vectordatasource.transform.normalize_operator_values
      - vectordatasource.transform.truncate_min_zoom_to_2dp
    sort: vectordatasource.sort.landuse
    area-inclusion-threshold: 1
  roads:
    template: roads.jinja2
    geometry_types: [LineString, MultiLineString]
    simplify_start: 8
    transform:
      - vectordatasource.transform.tags_create_dict
      - vectordatasource.transform.tags_name_i18n
      - vectordatasource.transform.add_id_to_properties
      - vectordatasource.transform.detect_osm_relation
      - vectordatasource.transform.parse_layer_as_float
      - vectordatasource.transform.road_classifier
      - vectordatasource.transform.road_oneway
      - vectordatasource.transform.road_abbreviate_name
      - vectordatasource.transform.route_name
      - vectordatasource.transform.normalize_aerialways
      - vectordatasource.transform.normalize_cycleway
      - vectordatasource.transform.add_is_bicycle_related
      - vectordatasource.transform.merge_networks_from_tags
      - vectordatasource.transform.extract_network_information
      - vectordatasource.transform.choose_most_important_network
      - vectordatasource.transform.road_trim_properties
      - vectordatasource.transform.remove_feature_id
      - vectordatasource.transform.tags_remove
      - vectordatasource.transform.truncate_min_zoom_to_2dp
    sort: vectordatasource.sort.roads
    area-inclusion-threshold: 1
  buildings:
    template: buildings.jinja2
    clip_factor: 3.0
    geometry_types: [Point, MultiPoint, Polygon, MultiPolygon]
    transform:
      - vectordatasource.transform.tags_create_dict
      - vectordatasource.transform.tags_name_i18n
      - vectordatasource.transform.tags_remove
      - vectordatasource.transform.add_id_to_properties
      - vectordatasource.transform.detect_osm_relation
      - vectordatasource.transform.parse_layer_as_float
      - vectordatasource.transform.building_height
      - vectordatasource.transform.building_min_height
      - vectordatasource.transform.synthesize_volume
      - vectordatasource.transform.normalize_tourism_kind
      - vectordatasource.transform.building_trim_properties
      - vectordatasource.transform.remove_feature_id
      - vectordatasource.transform.truncate_min_zoom_to_2dp
    sort: vectordatasource.sort.buildings
    area-inclusion-threshold: 1
  pois:
    template: pois.jinja2
    geometry_types: [Point, MultiPoint, Polygon, MultiPolygon]
    transform:
      - vectordatasource.transform.tags_create_dict
      - vectordatasource.transform.add_iata_code_to_airports
      - vectordatasource.transform.normalize_tourism_kind
      - vectordatasource.transform.normalize_social_kind
      - vectordatasource.transform.normalize_medical_kind
      - vectordatasource.transform.tags_name_i18n
      - vectordatasource.transform.add_uic_ref
      - vectordatasource.transform.tags_remove
      - vectordatasource.transform.add_id_to_properties
      - vectordatasource.transform.detect_osm_relation
      - vectordatasource.transform.remove_feature_id
      - vectordatasource.transform.remove_zero_area
      - vectordatasource.transform.make_representative_point
      - vectordatasource.transform.height_to_meters
      - vectordatasource.transform.pois_capacity_int
      - vectordatasource.transform.elevation_to_meters
      - vectordatasource.transform.normalize_operator_values
      - vectordatasource.transform.truncate_min_zoom_to_2dp
    sort: vectordatasource.sort.pois
    area-inclusion-threshold: 1
  boundaries:
    template: boundaries.jinja2
    geometry_types: [Polygon, MultiPolygon, LineString, MultiLineString]
    simplify_before_intersect: true
    simplify_start: 8
    transform:
      - vectordatasource.transform.tags_create_dict
      - vectordatasource.transform.tags_name_i18n
      - vectordatasource.transform.tags_remove
      - vectordatasource.transform.admin_level_as_int
      - vectordatasource.transform.add_id_to_properties
      - vectordatasource.transform.detect_osm_relation
      - vectordatasource.transform.remove_feature_id
      - vectordatasource.transform.truncate_min_zoom_to_2dp
    area-inclusion-threshold: 1
  transit:
    template: transit.jinja2
    geometry_types: [LineString, MultiLineString, Polygon, MultiPolygon]
    transform:
      - vectordatasource.transform.tags_create_dict
      - vectordatasource.transform.tags_name_i18n
      - vectordatasource.transform.add_uic_ref
      - vectordatasource.transform.tags_remove
      - vectordatasource.transform.add_id_to_properties
      - vectordatasource.transform.detect_osm_relation
      - vectordatasource.transform.route_name
      - vectordatasource.transform.parse_layer_as_float
      - vectordatasource.transform.remove_feature_id
      - vectordatasource.transform.truncate_min_zoom_to_2dp
    sort: vectordatasource.sort.transit
    area-inclusion-threshold: 1
post_process:
  - fn: vectordatasource.transform.numeric_min_filter
    params:
      source_layer: buildings
      mode: any
      filters:
        13: { area: 1600, volume: 300000 }
        14: { area:  100, volume: 200000 }
        15: { area:  100, volume: 100000 }

  - fn: vectordatasource.transform.build_fence
    params:
      base_layer: landuse
      start_zoom: 16
      prop_transform:
        kind: true
        id: true
        source: true
        area: true
        min_zoom: true
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: landuse
      start_zoom: 4
      properties: [barrier]

  # sort key
  - fn: vectordatasource.transform.csv_match_properties
    resources:
      matcher:
        type: file
        init_fn: vectordatasource.transform.CSVMatcher
        path: spreadsheets/sort_rank/landuse.csv
    params:
      source_layer: landuse
      target_value_type: int
  - fn: vectordatasource.transform.csv_match_properties
    resources:
      matcher:
        type: file
        init_fn: vectordatasource.transform.CSVMatcher
        path: spreadsheets/sort_rank/roads.csv
    params:
      source_layer: roads
      target_value_type: int
  - fn: vectordatasource.transform.csv_match_properties
    resources:
      matcher:
        type: file
        init_fn: vectordatasource.transform.CSVMatcher
        path: spreadsheets/sort_rank/earth.csv
    params:
      source_layer: earth
      target_value_type: int
  - fn: vectordatasource.transform.csv_match_properties
    resources:
      matcher:
        type: file
        init_fn: vectordatasource.transform.CSVMatcher
        path: spreadsheets/sort_rank/boundaries.csv
    params:
      source_layer: boundaries
      target_value_type: int
  - fn: vectordatasource.transform.csv_match_properties
    resources:
      matcher:
        type: file
        init_fn: vectordatasource.transform.CSVMatcher
        path: spreadsheets/sort_rank/buildings.csv
    params:
      source_layer: buildings
      target_value_type: int
  - fn: vectordatasource.transform.csv_match_properties
    resources:
      matcher:
        type: file
        init_fn: vectordatasource.transform.CSVMatcher
        path: spreadsheets/sort_rank/transit.csv
    params:
      source_layer: transit
      target_value_type: int
  # this needs to run before the water sort_rank csv matcher
  - fn: vectordatasource.transform.exterior_boundaries
    params:
      base_layer: water
      start_zoom: 9
      prop_transform:
        kind: true
        id: true
        source: true
        boundary: [true]
        area: true
      snap_tolerance: 0.125
  # have to do the water properties matching _after_ exterior boundaries
  # as it depends on having the "boundary: yes" property available.
  - fn: vectordatasource.transform.csv_match_properties
    resources:
      matcher:
        type: file
        init_fn: vectordatasource.transform.CSVMatcher
        path: spreadsheets/sort_rank/water.csv
    params:
      source_layer: water
      target_value_type: int

  # assign `scale_rank`
  - fn: vectordatasource.transform.csv_match_properties
    resources:
      matcher:
        type: file
        init_fn: vectordatasource.transform.CSVMatcher
        path: spreadsheets/scale_rank/buildings.csv
    params:
      source_layer: buildings
      target_value_type: int

  # drop all polygonal features that don't meet the minimum area
  # threshold. this depends on the mz_min_pixels property to be set on
  # features first
  - fn: vectordatasource.transform.drop_features_mz_min_pixels
    params:
      property: mz_min_pixels
      source_layers:
        # all layers that have polygonal geometries
        - boundaries
        - buildings
        - earth
        - landuse
        - transit
        - water

  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      properties: [layer]
  - fn: vectordatasource.transform.overlap
    params:
      base_layer: buildings
      cutting_layer: landuse
      attribute: kind
      target_attribute: landuse_kind
      cutting_attrs: { sort_key: 'sort_rank', reverse: True }
  - fn: vectordatasource.transform.admin_boundaries
    params:
      base_layer: boundaries
      start_zoom: 8

  # drop names on boundary lines (country, region, macroregion) except zoom 7 from Natural Earth
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: boundaries
      start_zoom: 0
      end_zoom: 7
      properties: [name]

  - fn: vectordatasource.transform.handle_label_placement
    params:
      layers:
        - water
        - earth
      location_property: mz_label_placement
      label_property_name: label_placement
      label_property_value: true
      label_where: >-
        'name' in properties

  - fn: vectordatasource.transform.handle_label_placement
    params:
      layers:
        - landuse
      start_zoom: 15
      location_property: mz_label_placement
      label_property_name: label_placement
      label_property_value: true
      label_where: >-
        'name' in properties or
        ('sport' in properties and properties.get('kind', 'rock') not in ('rock', 'stone'))

  # drop mz_label placement at zooms we don't apply handle_label_placement
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: landuse
      start_zoom: 0
      end_zoom: 15
      properties: [mz_label_placement]
      geom_types: [Polygon, MultiPolygon]

  - fn: vectordatasource.transform.handle_label_placement
    params:
      layers:
        - buildings
      start_zoom: 16
      location_property: mz_label_placement
      label_property_name: label_placement
      label_property_value: true
      label_where: >-
        'name' in properties

  # drop mz_label placement at zooms we don't apply handle_label_placement
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: buildings
      start_zoom: 0
      end_zoom: 16
      properties: [mz_label_placement]
      geom_types: [Polygon, MultiPolygon]

  - fn: vectordatasource.transform.remove_duplicate_features
    params:
      source_layer: landuse
      end_zoom: 16
      property_keys: [name, kind]
      geometry_types: [Point]
      min_distance: 256.0

  # drop the names and less important additional tags of small landuse polygons
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: landuse
      start_zoom: 4
      end_zoom: 13
      properties: [name, sport, religion, surface]
      where: >-
        pixel_area > area
      geom_types: [Polygon, MultiPolygon]

  - fn: vectordatasource.transform.drop_features_where
    params:
      source_layer: water
      start_zoom: 3
      where: >-
        'label_placement' not in properties and kind in ('sea', 'bay', 'strait', 'fjord')
  - fn: vectordatasource.transform.drop_features_where
    params:
      source_layer: earth
      start_zoom: 0
      where: >-
        'label_placement' not in properties and kind in
        ('archipelago', 'island', 'islet', 'continent', 'valley', 'ridge')

  # merge water and earth polygons, see https://github.com/tilezen/vector-datasource/issues/1106
  - fn: vectordatasource.transform.merge_polygon_features
    params:
      source_layer: earth
  - fn: vectordatasource.transform.merge_polygon_features
    params:
      source_layer: water

  - fn: vectordatasource.transform.generate_address_points
    params:
      source_layer: buildings
      start_zoom: 16
  - fn: vectordatasource.transform.remove_duplicate_features
    params:
      source_layer: buildings
      end_zoom: 16
      property_keys: [name, kind]
      geometry_types: [Point]
      min_distance: 256.0
  - fn: vectordatasource.transform.remove_duplicate_features
    params:
      source_layer: pois
      end_zoom: 16
      property_keys: [name, kind]
      geometry_types: [Point]
      min_distance: 256.0
  - fn: vectordatasource.transform.drop_features_where
    params:
      source_layer: landuse
      start_zoom: 6
      where: >-
        properties.get('mz_is_building') is True and
        properties.get('label_placement') is None
  - fn: vectordatasource.transform.merge_duplicate_stations
    params:
      source_layer: pois
      end_zoom: 16
  - fn: vectordatasource.transform.normalize_station_properties
    params:
      source_layer: pois
  - fn: vectordatasource.transform.rank_features
    params:
      source_layer: pois
      items_matching:
        kind: station
      rank_key: kind_tile_rank
  - fn: vectordatasource.transform.rank_features
    params:
      source_layer: places
      items_matching:
        kind: [neighbourhood, microhood, macrohood]
      rank_key: kind_tile_rank
      start_zoom: 11
  - fn: vectordatasource.transform.rank_features
    params:
      source_layer: pois
      items_matching:
        kind: [peak, volcano]
      rank_key: kind_tile_rank
      start_zoom: 9
  - fn: vectordatasource.transform.remove_duplicate_features
    params:
      source_layers: [pois, landuse, buildings]
      end_zoom: 16
      property_keys: [name]
      geometry_types: [Point]
      min_distance: 64.0
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 0
      end_zoom: 15
      properties: [is_bridge, is_tunnel]
      where: >-
        (kind == 'path' and zoom < 15) or
        (kind in ['minor_road', 'major_road', 'highway', 'rail'] and zoom < 13)
  # drop certain road properties at lower zooms
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 0
      end_zoom: 12
      properties:
        - bicycle
        - colour
        - foot
        - horse
        - motor_vehicle
        - operator
        - route          # NVK (2017425): this is a suspicious tag
        - route_name     # NVK (2017425): this is a suspicious tag
        - sac_scale
        - state
        - symbol
        - tracktype
        - type           # NVK (2017425): this is a suspicious tag
  # drop certain road properties at lower zooms
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 0
      end_zoom: 14
      properties:
        - all_bicycle_networks
        - all_bicycle_shield_texts
        - all_walking_networks
        - all_walking_shield_texts
        - all_bus_networks
        - all_bus_shield_texts
  # drop certain road properties at lower zooms
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 0
      end_zoom: 15
      properties:
        - ascent
        - crossing
        - descent
        - description
        - distance
        - incline
        - oneway
        - oneway_bicycle
        - ramp
        - ramp_bicycle
        - roundtrip
        - segregated
        - sidewalk
        - sidewalk_left
        - sidewalk_right
        - sport
        - trail_visibility
  # drop name and other properties on early paths to
  # allow more line merging (use walking network for labeling)
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 0
      end_zoom: 12
      properties:
        - name
        - ref
      where: >-
        kind == 'path'
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 0
      end_zoom: 15
      properties:
        - name
        - ref
        - network
        - shield_text
        - all_networks
        - all_shield_texts
        - bus_network
        - bus_shield_text
      where: >-
        (kind == 'rail' and zoom < 15) or
        (kind == 'minor_road' and zoom < 14) or
        (kind == 'major_road' and zoom <  7)
  # this is a patch because we still want to drop name, network from major_road
  # features between zoom 7 and 10
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 7
      end_zoom: 11
      properties:
        - name
        - all_networks
        - all_shield_texts
      where: >-
        kind == 'major_road'
  # this is a patch to get rid of name, but keep ref & network, for highways
  # when zoom < 11.
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 7
      end_zoom: 11
      properties: [name]
      where: >-
        kind == 'highway'
  # drop name, ref and the multi-shield properties, but keep single-shield
  # properties, when z<=6 (i.e: NE data)
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 0
      end_zoom: 7
      properties:
        - name
        - ref
        - all_networks
        - all_shield_texts
        - walking_network
        - walking_shield_text
        - bicycle_network
        - bicycle_shield_text
      where: >-
        kind == 'highway'
  # drop non-road shield stuff a good deal earlier.
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 0
      end_zoom: 11
      properties:
        - all_bicycle_networks
        - all_bicycle_shield_texts
        - all_walking_networks
        - all_walking_shield_texts
        - bus_network
        - bus_shield_text
        - all_bus_networks
        - all_bus_shield_texts
      where: >-
        kind == 'highway'
  # drop walking network related properties on all roads early
  # as paths are only brought in at zoom 9 and these properties
  # on other kinds would prevent merging at zoom 8 and earlier
  - fn: vectordatasource.transform.drop_properties
    params:
      source_layer: roads
      start_zoom: 0
      end_zoom: 9
      properties:
        - walking_network
        - walking_shield_text
  - fn: vectordatasource.transform.update_parenthetical_properties
    params:
      source_layer: pois
      values: ['closed', 'historical']
      target_min_zoom: 17
      drop_below_zoom: 16
  - fn: vectordatasource.transform.update_parenthetical_properties
    params:
      source_layer: buildings
      values: ['closed', 'historical']
      target_min_zoom: 17
      drop_below_zoom: 16
  - fn: vectordatasource.transform.merge_line_features
    params:
      source_layer: transit
      start_zoom: 0
      end_zoom: 16
  # only keep all stations below z12, then cut it down to the top few.
  - fn: vectordatasource.transform.keep_n_features
    params:
      source_layer: transit
      start_zoom: 9
      end_zoom: 12
      items_matching: { kind: station }
      max_items: 20
  - fn: vectordatasource.transform.keep_n_features
    params:
      source_layer: transit
      start_zoom: 12
      end_zoom: 14
      items_matching: { kind: station }
      max_items: 30
  - fn: vectordatasource.transform.keep_n_features
    params:
      source_layer: pois
      start_zoom: 9
      end_zoom: 16
      items_matching: { kind: [peak, volcano] }
      max_items: 5
  - fn: vectordatasource.transform.drop_properties_with_prefix
    params: {prefix: mz_}

  - fn: vectordatasource.transform.merge_building_features
    params:
      source_layer: buildings
      start_zoom: 13
      end_zoom: 15
      quantize:
        13: vectordatasource.transform.quantize_height_round_nearest_10_meters
        14: vectordatasource.transform.quantize_height_round_nearest_5_meters
      drop:
        - name
        - addr_housenumber
        - addr_street
      max_merged_features: 1000

  - fn: vectordatasource.transform.drop_small_inners
    params:
      end_zoom: 14
      source_layers: [landuse, water]
      pixel_area: 0.1

  - fn: vectordatasource.transform.simplify_and_clip
    params: {simplify_before: 16}

  - fn: vectordatasource.transform.intercut
    params:
      base_layer: roads
      cutting_layer: landuse
      attribute: kind
      target_attribute: landuse_kind
      cutting_attrs: { sort_key: 'sort_rank', reverse: True }
  # merge any remaining polygons by their properties.
  - fn: vectordatasource.transform.merge_polygon_features
    params:
      source_layer: landuse
      start_zoom: 4
      end_zoom: 13

  - fn: vectordatasource.transform.merge_line_features
    params:
      source_layer: roads
      start_zoom: 8
      end_zoom: 16

  - fn: vectordatasource.transform.simplify_layer
    params:
      source_layer: roads
      start_zoom: 8
      end_zoom: 16
      tolerance: 1.0

  - fn: vectordatasource.transform.buildings_unify
    params:
      source_layer: buildings
      start_zoom: 15

  - fn: vectordatasource.transform.palettize_colours
    params:
      layer: transit
      attribute: colour_name
      colours:
        aqua: [0, 255, 255]
        aquamarine: [127, 255, 212]
        black: [0, 0, 0]
        blue: [0, 0, 255]
        brown: [165, 42, 42]
        crimson: [220, 20, 60]
        darkgrey: [64, 64, 64]
        darkorchid: [153, 50, 204]
        darkred: [139, 0, 0]
        darkseagreen: [143, 188, 143]
        dodgerblue: [30, 144, 255]
        fuchsia: [255, 0, 255]
        gainsboro: [220, 220, 220]
        gold: [255, 215, 0]
        goldenrod: [218, 165, 32]
        green: [0, 128, 0]
        grey: [128, 128, 128]
        hotpink: [255, 105, 180]
        indigo: [75, 0, 130]
        khaki: [240, 230, 140]
        lightblue: [173, 216, 230]
        lightcoral: [240, 128, 128]
        lightgreen: [144, 238, 144]
        lightteal: [0, 153, 170]
        lime: [0, 255, 0]
        limegreen: [50, 205, 50]
        mediumpurple: [147, 112, 219]
        mediumseagreen: [60, 179, 113]
        mediumturquoise: [72, 209, 204]
        navy: [0, 0, 128]
        olivedrab: [107, 142, 35]
        orange: [255, 165, 0]
        orangered: [255, 69, 0]
        peru: [205, 133, 63]
        pink: [255, 192, 203]
        plum: [221, 160, 221]
        purple: [128, 0, 128]
        red: [255, 0, 0]
        royalblue: [65, 105, 225]
        sandybrown: [244, 164, 96]
        silver: [192, 192, 192]
        steelblue: [70, 130, 180]
        tan: [210, 180, 140]
        teal: [0, 128, 128]
        tomato: [255, 99, 71]
        violet: [238, 130, 238]
        white: [255, 255, 255]
        yellow: [255, 255, 0]
        yellowgreen: [154, 205, 50]