Implement FlatGeobuf Polygon type parsing [#2]

This commit is contained in:
Brandon Liu 2022-03-25 11:02:01 +08:00
parent a3dec5116d
commit 8a5a2496c3
4 changed files with 186 additions and 3 deletions

View File

@ -47,7 +47,7 @@ C = $(wildcard *.c) $(wildcard *.cpp)
INCLUDES = -I/usr/local/include -I.
LIBS = -L/usr/local/lib
tippecanoe: geojson.o jsonpull/jsonpull.o tile.o pool.o mbtiles.o geometry.o projection.o memfile.o mvt.o serial.o main.o text.o dirtiles.o plugin.o read_json.o write_json.o geobuf.o evaluator.o geocsv.o csv.o geojson-loop.o
tippecanoe: geojson.o jsonpull/jsonpull.o tile.o pool.o mbtiles.o geometry.o projection.o memfile.o mvt.o serial.o main.o text.o dirtiles.o plugin.o read_json.o write_json.o geobuf.o flatgeobuf.o evaluator.o geocsv.o csv.o geojson-loop.o
$(CXX) $(PG) $(LIBS) $(FINAL_FLAGS) $(CXXFLAGS) -o $@ $^ $(LDFLAGS) -lm -lz -lsqlite3 -lpthread
tippecanoe-enumerate: enumerate.o

109
flatgeobuf.cpp Normal file
View File

@ -0,0 +1,109 @@
#include <stdio.h>
#include "serial.hpp"
#include <iostream>
#include "projection.hpp"
#include "flatgeobuf/feature_generated.h"
#include "flatgeobuf/header_generated.h"
using namespace std;
struct NodeItem {
double minX;
double minY;
double maxX;
double maxY;
uint64_t offset;
};
// copied from https://github.com/flatgeobuf/flatgeobuf/blob/master/src/cpp/packedrtree.cpp#L365
uint64_t PackedRTreeSize(const uint64_t numItems, const uint16_t nodeSize)
{
if (nodeSize < 2)
throw std::invalid_argument("Node size must be at least 2");
if (numItems == 0)
throw std::invalid_argument("Number of items must be greater than 0");
const uint16_t nodeSizeMin = std::min(std::max(nodeSize, static_cast<uint16_t>(2)), static_cast<uint16_t>(65535));
// limit so that resulting size in bytes can be represented by uint64_t
if (numItems > static_cast<uint64_t>(1) << 56)
throw std::overflow_error("Number of items must be less than 2^56");
uint64_t n = numItems;
uint64_t numNodes = n;
do {
n = (n + nodeSizeMin - 1) / nodeSizeMin;
numNodes += n;
} while (n != 1);
return numNodes * sizeof(NodeItem);
}
drawvec readGeometry(const FlatGeobuf::Geometry *geometry, FlatGeobuf::GeometryType geometry_type) {
// if it is a GeometryCollection, parse Parts, ignore XY
auto xy = geometry->xy();
auto ends = geometry->ends();
size_t current_end = 0;
drawvec dv;
for (unsigned int i = 0; i < xy->size(); i+=2) {
long long x, y;
projection->project(xy->Get(i), xy->Get(i+1), 32, &x, &y);
if (i == 0 || (ends != nullptr && current_end < ends->size() && i == ends->Get(current_end))) {
dv.push_back(draw(VT_MOVETO, x, y));
if (i > 0) current_end++;
} else {
dv.push_back(draw(VT_LINETO, x, y));
}
}
return dv;
}
void parse_flatgeobuf(std::vector<struct serialization_state> *sst, const char *src, int layer, std::string layername) {
auto header_size = flatbuffers::GetPrefixedSize((const uint8_t *)src + 8);
auto header = FlatGeobuf::GetSizePrefixedHeader(src + 8);
auto features_count = header->features_count();
auto node_size = header->index_node_size();
auto index_size = PackedRTreeSize(features_count,node_size);
auto h_geometry_type = header->geometry_type();
switch (h_geometry_type) {
case FlatGeobuf::GeometryType::Unknown :
case FlatGeobuf::GeometryType::Point :
case FlatGeobuf::GeometryType::LineString :
case FlatGeobuf::GeometryType::Polygon :
case FlatGeobuf::GeometryType::MultiPoint :
case FlatGeobuf::GeometryType::MultiLineString :
case FlatGeobuf::GeometryType::MultiPolygon :
case FlatGeobuf::GeometryType::GeometryCollection :
break;
default:
fprintf(stderr, "flatgeobuf has unsupported geometry type\n");
exit(EXIT_FAILURE);
}
const char* start = src + 8 + 4 + header_size + index_size;
for (size_t i = 0; i < features_count; i++) {
serial_feature sf;
auto my_sst = &(*sst)[0];
auto feature_size = flatbuffers::GetPrefixedSize((const uint8_t *)start);
auto feature = FlatGeobuf::GetSizePrefixedFeature(start);
drawvec dv = readGeometry(feature->geometry(), h_geometry_type);
sf.layer = layer;
sf.layername = layername;
sf.segment = my_sst->segment;
sf.has_id = false;
sf.has_tippecanoe_minzoom = false;
sf.has_tippecanoe_maxzoom = false;
sf.feature_minzoom = false;
sf.seq = (*my_sst->layer_seq);
sf.geometry = dv;
sf.t = 4;
serialize_feature(my_sst, sf);
start += 4 + feature_size;
}
}

View File

@ -1,2 +1,6 @@
#include "flatgeobuf/feature_generated.h"
#include "flatgeobuf/header_generated.h"
#ifndef FLATGEOBUF_HPP
#define FLATGEOBUF_HPP
void parse_flatgeobuf(std::vector<struct serialization_state> *sst, const char *s, int layer, std::string layername);
#endif

View File

@ -1354,6 +1354,76 @@ int read_input(std::vector<source> &sources, char *fname, int maxzoom, int minzo
}
size_t layer = a->second.id;
// geobuf
if (sources[source].format == "fgb" || (sources[source].file.size() > 4 && sources[source].file.substr(sources[source].file.size() - 4) == std::string(".fgb"))) {
struct stat st;
if (fstat(fd, &st) != 0) {
perror("fstat");
perror(sources[source].file.c_str());
exit(EXIT_FAILURE);
}
char *map = (char *) mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
if (map == MAP_FAILED) {
fprintf(stderr, "%s: mmap: %s: %s\n", *av, reading.c_str(), strerror(errno));
exit(EXIT_FAILURE);
}
std::atomic<long long> layer_seq[CPUS];
double dist_sums[CPUS];
size_t dist_counts[CPUS];
std::vector<struct serialization_state> sst;
sst.resize(CPUS);
for (size_t i = 0; i < CPUS; i++) {
layer_seq[i] = overall_offset;
dist_sums[i] = 0;
dist_counts[i] = 0;
sst[i].fname = reading.c_str();
sst[i].line = 0;
sst[i].layer_seq = &layer_seq[i];
sst[i].progress_seq = &progress_seq;
sst[i].readers = &readers;
sst[i].segment = i;
sst[i].initial_x = &initial_x[i];
sst[i].initial_y = &initial_y[i];
sst[i].initialized = &initialized[i];
sst[i].dist_sum = &dist_sums[i];
sst[i].dist_count = &dist_counts[i];
sst[i].want_dist = guess_maxzoom;
sst[i].maxzoom = maxzoom;
sst[i].filters = prefilter != NULL || postfilter != NULL;
sst[i].uses_gamma = uses_gamma;
sst[i].layermap = &layermaps[i];
sst[i].exclude = exclude;
sst[i].include = include;
sst[i].exclude_all = exclude_all;
sst[i].basezoom = basezoom;
sst[i].attribute_types = attribute_types;
}
parse_flatgeobuf(&sst, map, st.st_size, layer, sources[layer].layer);
for (size_t i = 0; i < CPUS; i++) {
dist_sum += dist_sums[i];
dist_count += dist_counts[i];
}
if (munmap(map, st.st_size) != 0) {
perror("munmap source file");
exit(EXIT_FAILURE);
}
if (close(fd) != 0) {
perror("close");
exit(EXIT_FAILURE);
}
overall_offset = layer_seq[0];
checkdisk(&readers);
continue;
}
if (sources[source].format == "geobuf" || (sources[source].file.size() > 7 && sources[source].file.substr(sources[source].file.size() - 7) == std::string(".geobuf"))) {
struct stat st;
if (fstat(fd, &st) != 0) {