1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
use std::collections::{BTreeMap, HashSet};

use rand::{Rng, SeedableRng};
use rand_xorshift::XorShiftRng;

use abstutil::{Tags, Timer};
use geom::{Distance, HashablePt2D, Line, Polygon};

use crate::make::match_points_to_lanes;
use crate::raw::RawBuilding;
use crate::{
    osm, Amenity, Building, BuildingID, BuildingType, LaneID, Map, NamePerLanguage,
    OffstreetParking,
};

/// Finalize importing of buildings, mostly by matching them to the nearest sidewalk.
pub fn make_all_buildings(
    input: &BTreeMap<osm::OsmID, RawBuilding>,
    map: &Map,
    keep_bldg_tags: bool,
    timer: &mut Timer,
) -> Vec<Building> {
    timer.start("convert buildings");
    let mut center_per_bldg: BTreeMap<osm::OsmID, HashablePt2D> = BTreeMap::new();
    let mut query: HashSet<HashablePt2D> = HashSet::new();
    timer.start_iter("get building center points", input.len());
    for (id, b) in input {
        timer.next();
        let center = b.polygon.center().to_hashable();
        center_per_bldg.insert(*id, center);
        query.insert(center);
    }

    // equiv_pos could be a little closer, so use two buffers
    let sidewalk_buffer = Distance::meters(7.5);
    let sidewalk_pts = match_points_to_lanes(
        map.get_bounds(),
        query,
        map.all_lanes(),
        |l| l.is_walkable(),
        // Don't put connections too close to intersections
        sidewalk_buffer,
        // Try not to skip any buildings, but more than 1km from a sidewalk is a little much
        Distance::meters(1000.0),
        timer,
    );

    let mut results = Vec::new();
    timer.start_iter("match buildings to sidewalks", center_per_bldg.len());
    for (orig_id, bldg_center) in center_per_bldg {
        timer.next();
        if let Some(sidewalk_pos) = sidewalk_pts.get(&bldg_center) {
            let b = &input[&orig_id];
            let sidewalk_line = match Line::new(bldg_center.to_pt2d(), sidewalk_pos.pt(map)) {
                Some(l) => trim_path(&b.polygon, l),
                None => {
                    warn!(
                        "Skipping building {} because front path has 0 length",
                        orig_id
                    );
                    continue;
                }
            };

            let id = BuildingID(results.len());

            let mut rng = XorShiftRng::seed_from_u64(orig_id.inner() as u64);
            // TODO is it worth using height or building:height as an alternative if not tagged?
            let levels = b
                .osm_tags
                .get("building:levels")
                .and_then(|x| x.parse::<f64>().ok())
                .unwrap_or(1.0);

            results.push(Building {
                id,
                polygon: b.polygon.clone(),
                levels,
                address: get_address(&b.osm_tags, sidewalk_pos.lane(), map),
                name: NamePerLanguage::new(&b.osm_tags),
                orig_id,
                label_center: b.polygon.polylabel(),
                amenities: if keep_bldg_tags {
                    b.amenities.clone()
                } else {
                    b.amenities
                        .iter()
                        .map(|a| {
                            let mut a = a.clone();
                            a.osm_tags = Tags::empty();
                            a
                        })
                        .collect()
                },
                bldg_type: classify_bldg(
                    &b.osm_tags,
                    &b.amenities,
                    levels,
                    b.polygon.area(),
                    &mut rng,
                ),
                parking: if let Some(n) = b.public_garage_name.clone() {
                    OffstreetParking::PublicGarage(n, b.num_parking_spots)
                } else {
                    OffstreetParking::Private(
                        b.num_parking_spots,
                        b.osm_tags.is("building", "parking") || b.osm_tags.is("amenity", "parking"),
                    )
                },
                osm_tags: if keep_bldg_tags {
                    b.osm_tags.clone()
                } else {
                    Tags::empty()
                },

                sidewalk_pos: *sidewalk_pos,
                driveway_geom: sidewalk_line.to_polyline(),
            });
        }
    }

    info!(
        "Discarded {} buildings that weren't close enough to a sidewalk",
        input.len() - results.len()
    );
    timer.stop("convert buildings");

    results
}

// Adjust the path to start on the building's border, not center
fn trim_path(poly: &Polygon, path: Line) -> Line {
    for bldg_line in poly.points().windows(2) {
        if let Some(l1) = Line::new(bldg_line[0], bldg_line[1]) {
            if let Some(hit) = l1.intersection(&path) {
                if let Some(l2) = Line::new(hit, path.pt2()) {
                    return l2;
                }
            }
        }
    }
    // Just give up
    path
}

fn get_address(tags: &Tags, sidewalk: LaneID, map: &Map) -> String {
    match (tags.get("addr:housenumber"), tags.get("addr:street")) {
        (Some(num), Some(st)) => format!("{} {}", num, st),
        (None, Some(st)) => format!("??? {}", st),
        _ => format!("??? {}", map.get_parent(sidewalk).get_name(None)),
    }
}

fn classify_bldg(
    tags: &Tags,
    amenities: &Vec<Amenity>,
    levels: f64,
    ground_area_sq_meters: f64,
    rng: &mut XorShiftRng,
) -> BuildingType {
    // used: top values from https://taginfo.openstreetmap.org/keys/building#values (>100k uses)

    let mut commercial = false;

    let area_sq_meters = levels * ground_area_sq_meters;

    // These are produced by get_bldg_amenities in convert_osm/src/osm_reader.rs.
    // TODO: is it safe to assume all amenities are commercial?
    // TODO: consider converting amenities to an enum - maybe with a catchall case for the long
    //       tail of rarely used enums.
    if !amenities.is_empty() {
        commercial = true;
    }

    if tags.is("ruins", "yes") {
        if commercial {
            return BuildingType::Commercial(0);
        }
        return BuildingType::Empty;
    }

    let mut residents: usize = 0;
    let mut workers: usize = 0;

    if tags.is_any(
        "building",
        vec![
            "office",
            "industrial",
            "commercial",
            "retail",
            "warehouse",
            "civic",
            "public",
        ],
    ) {
        // 1 person per 10 square meters
        // TODO: Hone in this parameter. Space per person varies with (among other things):
        //  - building type. e.g. office vs. warehouse
        //  - regional/cultural norms
        workers = (area_sq_meters / 10.0) as usize;
    } else if tags.is_any(
        "building",
        vec!["school", "university", "construction", "church"],
    ) {
        // TODO: special handling in future
        return BuildingType::Empty;
    } else if tags.is_any(
        "building",
        vec![
            "garage",
            "garages",
            "shed",
            "roof",
            "greenhouse",
            "farm_auxiliary",
            "barn",
            "service",
        ],
    ) {
        return BuildingType::Empty;
    } else if tags.is_any(
        "building",
        vec!["house", "detached", "semidetached_house", "farm"],
    ) {
        residents = rng.gen_range(0..3);
    } else if tags.is_any("building", vec!["hut", "static_caravan", "cabin"]) {
        residents = rng.gen_range(0..2);
    } else if tags.is_any("building", vec!["apartments", "terrace", "residential"]) {
        // 1 person per 10 square meters
        // TODO: Hone in this parameter. Space per person varies with (among other things):
        //  - building type. e.g. apartment vs single family
        //  - regional/cultural norms
        residents = (area_sq_meters / 10.0) as usize;
    } else {
        residents = rng.gen_range(0..2);
    }

    if commercial && workers == 0 {
        // TODO: Come up with a better measure
        workers = (residents as f64 / 3.0) as usize;
    }

    if commercial {
        if residents > 0 {
            return BuildingType::ResidentialCommercial(residents, workers);
        }
        return BuildingType::Commercial(workers);
    }
    return BuildingType::Residential {
        num_residents: residents,
        num_housing_units: 1,
    };
}