tippecanoe/decode.cpp

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sqlite3.h>
#include <getopt.h>
#include <string>
#include <vector>
#include <map>
#include <set>
#include <zlib.h>
#include <math.h>
#include <fcntl.h>
#include <dirent.h>
#include <sys/stat.h>
#include <sys/mman.h>
#include <protozero/pbf_reader.hpp>
#include "mvt.hpp"
#include "projection.hpp"
#include "geometry.hpp"
#include "write_json.hpp"
#include "jsonpull/jsonpull.h"
#include "mbtiles.hpp"
#include "dirtiles.hpp"
#include "pmtiles_file.hpp"
#include "errors.hpp"
#include "usage.hpp"

int minzoom = 0;
int maxzoom = 32;
bool force = false;
std::set<std::string> include_attr;

bool progress_time() {
	return false;
}

void do_stats(mvt_tile &tile, size_t size, bool compressed, int z, unsigned x, unsigned y, json_writer &state) {
	state.json_write_hash();

	state.json_write_string("zoom");
	state.json_write_signed(z);

	state.json_write_string("x");
	state.json_write_unsigned(x);

	state.json_write_string("y");
	state.json_write_unsigned(y);

	state.json_write_string("bytes");
	state.json_write_unsigned(size);

	state.json_write_string("compressed");
	state.json_write_bool(compressed);

	state.json_write_string("layers");
	state.json_write_hash();

	for (size_t i = 0; i < tile.layers.size(); i++) {
		state.json_write_string(tile.layers[i].name);

		size_t points = 0, lines = 0, polygons = 0;
		for (size_t j = 0; j < tile.layers[i].features.size(); j++) {
			if (tile.layers[i].features[j].type == mvt_point) {
				points++;
			} else if (tile.layers[i].features[j].type == mvt_linestring) {
				lines++;
			} else if (tile.layers[i].features[j].type == mvt_polygon) {
				polygons++;
			}
		}

		state.json_write_hash();

		state.json_write_string("points");
		state.json_write_unsigned(points);

		state.json_write_string("lines");
		state.json_write_unsigned(lines);

		state.json_write_string("polygons");
		state.json_write_unsigned(polygons);

		state.json_write_string("extent");
		state.json_write_signed(tile.layers[i].extent);

		state.json_end_hash();
	}

	state.json_end_hash();
	state.json_end_hash();

	state.json_write_newline();
}

void handle(std::string message, int z, unsigned x, unsigned y, std::set<std::string> const &to_decode, bool pipeline, bool stats, json_writer &state, int coordinate_mode) {
	mvt_tile tile;
	bool was_compressed;

	try {
		if (!tile.decode(message, was_compressed)) {
			fprintf(stderr, "Couldn't parse tile %d/%u/%u\n", z, x, y);
			exit(EXIT_MVT);
		}
	} catch (std::exception const &e) {
		fprintf(stderr, "PBF decoding error in tile %d/%u/%u\n", z, x, y);
		exit(EXIT_PROTOBUF);
	}

	if (stats) {
		do_stats(tile, message.size(), was_compressed, z, x, y, state);
		return;
	}

	if (!pipeline) {
		state.json_write_hash();

		state.json_write_string("type");
		state.json_write_string("FeatureCollection");

		if (true) {
			state.json_write_string("properties");
			state.json_write_hash();

			state.json_write_string("zoom");
			state.json_write_signed(z);

			state.json_write_string("x");
			state.json_write_signed(x);

			state.json_write_string("y");
			state.json_write_signed(y);

			if (!was_compressed) {
				state.json_write_string("compressed");
				state.json_write_bool(false);
			}

			state.json_end_hash();

			if (projection != projections) {
				state.json_write_string("crs");
				state.json_write_hash();

				state.json_write_string("type");
				state.json_write_string("name");

				state.json_write_string("properties");
				state.json_write_hash();

				state.json_write_string("name");
				state.json_write_string(projection->alias);

				state.json_end_hash();
				state.json_end_hash();
			}
		}

		state.json_write_string("features");
		state.json_write_array();
		state.json_write_newline();
	}

	bool first_layer = true;
	for (size_t l = 0; l < tile.layers.size(); l++) {
		mvt_layer &layer = tile.layers[l];

		if (layer.extent <= 0) {
			fprintf(stderr, "Impossible layer extent %lld in mbtiles\n", layer.extent);
			exit(EXIT_IMPOSSIBLE);
		}

		if (to_decode.size() != 0 && !to_decode.count(layer.name)) {
			continue;
		}

		if (!pipeline) {
			if (true) {
				if (!first_layer) {
					state.json_comma_newline();
				}

				state.json_write_hash();

				state.json_write_string("type");
				state.json_write_string("FeatureCollection");

				state.json_write_string("properties");
				state.json_write_hash();

				state.json_write_string("layer");
				state.json_write_string(layer.name);

				state.json_write_string("version");
				state.json_write_signed(layer.version);

				state.json_write_string("extent");
				state.json_write_signed(layer.extent);

				state.json_end_hash();

				state.json_write_string("features");
				state.json_write_array();

				state.json_write_newline();
				first_layer = false;
			}
		}

		// X and Y are unsigned, so no need to check <0
		if (x > (1ULL << z) || y > (1ULL << z)) {
			fprintf(stderr, "Impossible tile %d/%u/%u\n", z, x, y);
			exit(EXIT_IMPOSSIBLE);
		}

		double scale = 0;
		if (coordinate_mode == 1) {  // fraction
			scale = layer.extent;
		} else if (coordinate_mode == 2) {  // integer
			scale = 1;
		}
		layer_to_geojson(layer, z, x, y, !pipeline, pipeline, pipeline, false, 0, 0, 0, !force, state, scale, include_attr);

		if (!pipeline) {
			if (true) {
				state.json_end_array();
				state.json_end_hash();
				state.json_write_newline();
			}
		}
	}

	if (!pipeline) {
		state.json_end_array();
		state.json_end_hash();
		state.json_write_newline();
	}
}

void decode(char *fname, int z, unsigned x, unsigned y, std::set<std::string> const &to_decode, bool pipeline, bool stats, std::set<std::string> const &exclude_meta, int coordinate_mode) {
	sqlite3 *db = NULL;
	bool isdir = false;
	int oz = z;
	unsigned ox = x, oy = y;
	json_writer state(stdout);

	int fd = open(fname, O_RDONLY | O_CLOEXEC);
	if (fd >= 0) {
		struct stat st;
		if (fstat(fd, &st) == 0) {
			char *map = (char *) mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
			if (map != NULL && map != MAP_FAILED) {
				if (strcmp(map, "SQLite format 3") != 0 && strncmp(map, "PMTiles", 7) != 0) {
					if (z >= 0) {
						if (st.st_size > 250 * 1024 * 1024) {
							fprintf(stderr, "%s: unrealistically large single-tile size %zu\n", fname, (size_t) st.st_size);
							exit(EXIT_MEMORY);
						} else {
							std::string s = std::string(map, st.st_size);
							handle(s, z, x, y, to_decode, pipeline, stats, state, coordinate_mode);
							munmap(map, st.st_size);
							return;
						}
					} else {
						fprintf(stderr, "Must specify zoom/x/y to decode a single pbf file\n");
						exit(EXIT_ARGS);
					}
				}
			}
			munmap(map, st.st_size);
		} else {
			perror("fstat");
		}
		if (close(fd) != 0) {
			perror("close");
			exit(EXIT_CLOSE);
		}
	} else {
		perror(fname);
	}

	struct stat st;
	std::vector<zxy> tiles;

	char *pmtiles_map = NULL;
	std::vector<pmtiles::entry_zxy> entries;
	bool is_pmtiles = false;

	if (stat(fname, &st) == 0 && (st.st_mode & S_IFDIR) != 0) {
		isdir = true;

		db = dirmeta2tmp(fname);
		tiles = enumerate_dirtiles(fname, minzoom, maxzoom);
	} else if (pmtiles_has_suffix(fname)) {
		int pmtiles_fd = open(fname, O_RDONLY | O_CLOEXEC);
		pmtiles_map = (char *) mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, pmtiles_fd, 0);
		if (pmtiles_map == MAP_FAILED) {
			perror("mmap in decode");
			exit(EXIT_MEMORY);
		}
		if (close(pmtiles_fd) != 0) {
			perror("close");
			exit(EXIT_CLOSE);
		}
		db = pmtilesmeta2tmp(fname, pmtiles_map);
		entries = pmtiles_entries_tms(pmtiles_map, minzoom, maxzoom);
		is_pmtiles = true;
	} else {
		if (sqlite3_open(fname, &db) != SQLITE_OK) {
			fprintf(stderr, "%s: %s\n", fname, sqlite3_errmsg(db));
			exit(EXIT_OPEN);
		}

		char *err = NULL;
		if (sqlite3_exec(db, "PRAGMA integrity_check;", NULL, NULL, &err) != SQLITE_OK) {
			fprintf(stderr, "%s: integrity_check: %s\n", fname, err);
			exit(EXIT_SQLITE);
		}
	}

	if (z < 0) {
		int within = 0;

		if (!pipeline && !stats) {
			state.json_write_hash();

			state.json_write_string("type");
			state.json_write_string("FeatureCollection");

			state.json_write_string("properties");
			state.json_write_hash();
			state.json_write_newline();

			const char *sql2 = "SELECT name, value from metadata order by name;";
			sqlite3_stmt *stmt2;
			if (sqlite3_prepare_v2(db, sql2, -1, &stmt2, NULL) != SQLITE_OK) {
				fprintf(stderr, "%s: select failed: %s\n", fname, sqlite3_errmsg(db));
				exit(EXIT_SQLITE);
			}

			while (sqlite3_step(stmt2) == SQLITE_ROW) {
				const unsigned char *name = sqlite3_column_text(stmt2, 0);
				const unsigned char *value = sqlite3_column_text(stmt2, 1);

				if (name == NULL || value == NULL) {
					fprintf(stderr, "Corrupt mbtiles file: null metadata\n");
					exit(EXIT_SQLITE);
				}

				if (exclude_meta.count((char *) name) == 0) {
					if (within) {
						state.json_comma_newline();
					}
					within = 1;

					state.json_write_string((char *) name);
					state.json_write_string((char *) value);
				}
			}

			state.json_write_newline();
			state.wantnl = false;  // XXX

			sqlite3_finalize(stmt2);
		}

		if (stats) {
			state.json_write_array();
			state.json_write_newline();
		}

		if (!pipeline && !stats) {
			state.json_end_hash();

			state.json_write_string("features");
			state.json_write_array();
			state.json_write_newline();
		}

		if (isdir) {
			within = 0;
			for (size_t i = 0; i < tiles.size(); i++) {
				if (!pipeline && !stats) {
					if (within) {
						state.json_comma_newline();
					}
					within = 1;
				}
				if (stats) {
					if (within) {
						state.json_comma_newline();
					}
					within = 1;
				}

				std::string fn = std::string(fname) + "/" + tiles[i].path();
				FILE *f = fopen(fn.c_str(), "rb");
				if (f == NULL) {
					perror(fn.c_str());
					exit(EXIT_OPEN);
				}

				std::string s;
				char buf[2000];
				ssize_t n;
				while ((n = fread(buf, 1, 2000, f)) > 0) {
					s.append(std::string(buf, n));
				}
				fclose(f);

				handle(s, tiles[i].z, tiles[i].x, tiles[i].y, to_decode, pipeline, stats, state, coordinate_mode);
			}
		} else if (is_pmtiles) {
			within = 0;

			for (auto const &entry : entries) {
				if (!pipeline && !stats) {
					if (within) {
						state.json_comma_newline();
					}
					within = 1;
				}
				if (stats) {
					if (within) {
						state.json_comma_newline();
					}
					within = 1;
				}

				std::string s{pmtiles_map + entry.offset, entry.length};
				handle(s, entry.z, entry.x, entry.y, to_decode, pipeline, stats, state, coordinate_mode);
			}
		} else {
			const char *sql = "SELECT tile_data, zoom_level, tile_column, tile_row from tiles where zoom_level between ? and ? order by zoom_level, tile_column, tile_row;";
			sqlite3_stmt *stmt;
			if (sqlite3_prepare_v2(db, sql, -1, &stmt, NULL) != SQLITE_OK) {
				fprintf(stderr, "%s: select failed: %s\n", fname, sqlite3_errmsg(db));
				exit(EXIT_SQLITE);
			}

			sqlite3_bind_int(stmt, 1, minzoom);
			sqlite3_bind_int(stmt, 2, maxzoom);

			within = 0;
			while (sqlite3_step(stmt) == SQLITE_ROW) {
				if (!pipeline && !stats) {
					if (within) {
						state.json_comma_newline();
					}
					within = 1;
				}
				if (stats) {
					if (within) {
						state.json_comma_newline();
					}
					within = 1;
				}

				int len = sqlite3_column_bytes(stmt, 0);
				int tz = sqlite3_column_int(stmt, 1);
				int tx = sqlite3_column_int(stmt, 2);
				int ty = sqlite3_column_int(stmt, 3);

				if (tz < 0 || tz >= 32) {
					fprintf(stderr, "Impossible zoom level %d in mbtiles\n", tz);
					exit(EXIT_IMPOSSIBLE);
				}

				ty = (1LL << tz) - 1 - ty;
				const char *s = (const char *) sqlite3_column_blob(stmt, 0);

				if (s == NULL) {
					fprintf(stderr, "Corrupt mbtiles file: null entry in tiles table\n");
					exit(EXIT_SQLITE);
				}

				handle(std::string(s, len), tz, tx, ty, to_decode, pipeline, stats, state, coordinate_mode);
			}

			sqlite3_finalize(stmt);
		}

		if (!pipeline && !stats) {
			state.json_end_array();
			state.json_end_hash();
			state.json_write_newline();
		}
		if (stats) {
			state.json_end_array();
			state.json_write_newline();
		}
		if (pipeline) {
			state.json_write_newline();
		}
	} else {
		int handled = 0;
		while (z >= 0 && !handled) {
			if (is_pmtiles) {
				uint64_t tile_offset;
				uint32_t tile_length;
				std::tie(tile_offset, tile_length) = pmtiles_get_tile(pmtiles_map, z, x, y);
				if (tile_length > 0) {
					if (z != oz) {
						fprintf(stderr, "%s: Warning: using tile %d/%u/%u instead of %d/%u/%u\n", fname, z, x, y, oz, ox, oy);
					}
					std::string s{pmtiles_map + tile_offset, tile_length};
					handle(s, z, x, y, to_decode, pipeline, stats, state, coordinate_mode);
					handled = 1;
				}
			} else {
				const char *sql = "SELECT tile_data from tiles where zoom_level = ? and tile_column = ? and tile_row = ?;";
				sqlite3_stmt *stmt;
				if (sqlite3_prepare_v2(db, sql, -1, &stmt, NULL) != SQLITE_OK) {
					fprintf(stderr, "%s: select failed: %s\n", fname, sqlite3_errmsg(db));
					exit(EXIT_SQLITE);
				}

				sqlite3_bind_int(stmt, 1, z);
				sqlite3_bind_int(stmt, 2, x);
				sqlite3_bind_int(stmt, 3, (1LL << z) - 1 - y);

				while (sqlite3_step(stmt) == SQLITE_ROW) {
					int len = sqlite3_column_bytes(stmt, 0);
					const char *s = (const char *) sqlite3_column_blob(stmt, 0);

					if (s == NULL) {
						fprintf(stderr, "Corrupt mbtiles file: null entry in tiles table\n");
						exit(EXIT_SQLITE);
					}

					if (z != oz) {
						fprintf(stderr, "%s: Warning: using tile %d/%u/%u instead of %d/%u/%u\n", fname, z, x, y, oz, ox, oy);
					}

					handle(std::string(s, len), z, x, y, to_decode, pipeline, stats, state, coordinate_mode);
					handled = 1;
				}
				sqlite3_finalize(stmt);
			}

			z--;
			x /= 2;
			y /= 2;
		}
	}

	if (sqlite3_close(db) != SQLITE_OK) {
		fprintf(stderr, "%s: could not close database: %s\n", fname, sqlite3_errmsg(db));
		exit(EXIT_CLOSE);
	}
}

static const struct option long_options[] = {
	{"Tiles to decode", 0, 0, 0},
	{"minimum-zoom", required_argument, 0, 'Z'},
	{"maximum-zoom", required_argument, 0, 'z'},
	{"layer", required_argument, 0, 'l'},

	{"Output format", 0, 0, 0},
	{"projection", required_argument, 0, 's'},
	{"fractional-coordinates", no_argument, 0, 'F'},
	{"integer-coordinates", no_argument, 0, 'I'},
	{"tag-layer-and-zoom", no_argument, 0, 'c'},
	{"stats", no_argument, 0, 'S'},

	{"Filtering the output", 0, 0, 0},
	{"include", required_argument, 0, 'y'},
	{"exclude-metadata-row", required_argument, 0, 'x'},

	{"Ignoring errors in the input", 0, 0, 0},
	{"force", no_argument, 0, 'f'},

	{0, 0, 0, 0},
};

// the options above, with the usage message headings removed
static struct option real_long_options[sizeof(long_options) / sizeof(long_options[0])];

void usage(char **argv) {
	static const char *const forms[] = {
		"[options] tileset",
		"[options] tileset zoom x y",
		NULL,
	};

	print_usage(stderr, argv[0], forms, long_options, NULL);
	fprintf(stderr, "\nThe tileset can be an .mbtiles or .pmtiles file or a directory of tiles,\n");
	fprintf(stderr, "or, if zoom/x/y is specified, a single .pbf tile.\n");
	exit(EXIT_ARGS);
}

int main(int argc, char **argv) {
	extern int optind;
	extern char *optarg;
	int i;
	std::set<std::string> to_decode;
	bool pipeline = false;
	bool stats = false;
	std::set<std::string> exclude_meta;
	int coordinate_mode = 0;

	strip_usage_headings(long_options, real_long_options);
	std::string getopt_str = getopt_string(real_long_options);

	while ((i = getopt_long(argc, argv, getopt_str.c_str(), real_long_options, NULL)) != -1) {
		switch (i) {
		case 0:
			break;

		case 's':
			set_projection_or_exit(optarg);
			break;

		case 'F':
			coordinate_mode = 1;
			break;

		case 'I':
			coordinate_mode = 2;
			break;

		case 'z':
			maxzoom = atoi(optarg);
			break;

		case 'Z':
			minzoom = atoi(optarg);
			break;

		case 'l':
			to_decode.insert(optarg);
			break;

		case 'c':
			pipeline = true;
			break;

		case 'S':
			stats = true;
			break;

		case 'f':
			force = true;
			break;

		case 'x':
			exclude_meta.insert(optarg);
			break;

		case 'y':
			include_attr.insert(optarg);
			break;

		default:
			usage(argv);
		}
	}

	if (argc == optind + 4) {
		decode(argv[optind], atoi(argv[optind + 1]), atoi(argv[optind + 2]), atoi(argv[optind + 3]), to_decode, pipeline, stats, exclude_meta, coordinate_mode);
	} else if (argc == optind + 1) {
		decode(argv[optind], -1, -1, -1, to_decode, pipeline, stats, exclude_meta, coordinate_mode);
	} else {
		usage(argv);
	}

	return 0;
}