Merge pull request #425 from mapbox/better-maxzoom-guessing

Better maxzoom guessing by considering resolution within features
This commit is contained in:
Eric Fischer
2017-05-30 17:22:40 -07:00
committed by GitHub
10 changed files with 2440 additions and 1015 deletions
+5
View File
@@ -1,3 +1,8 @@
## 1.19.0
* Tile-join can merge and create directories, not only mbtiles
* Maxzoom guessing (-zg) takes into account resolution within each feature
## 1.18.2
* Fix crash with very long (>128K) attribute values
+34 -6
View File
@@ -20,6 +20,7 @@
#include <sys/resource.h>
#include <pthread.h>
#include <vector>
#include <algorithm>
#include <set>
#include <map>
#include <string>
@@ -184,7 +185,7 @@ long long parse_geometry(int t, json_object *j, long long *bbox, drawvec &out, i
return g;
}
int serialize_geometry(json_object *geometry, json_object *properties, json_object *id, const char *reading, int line, volatile long long *layer_seq, volatile long long *progress_seq, long long *metapos, long long *geompos, long long *indexpos, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, FILE *metafile, FILE *geomfile, FILE *indexfile, struct memfile *poolfile, struct memfile *treefile, const char *fname, int basezoom, int layer, double droprate, long long *file_bbox, json_object *tippecanoe, int segment, int *initialized, unsigned *initial_x, unsigned *initial_y, struct reader *readers, int maxzoom, json_object *feature, std::map<std::string, layermap_entry> *layermap, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types) {
int serialize_geometry(json_object *geometry, json_object *properties, json_object *id, const char *reading, int line, volatile long long *layer_seq, volatile long long *progress_seq, long long *metapos, long long *geompos, long long *indexpos, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, FILE *metafile, FILE *geomfile, FILE *indexfile, struct memfile *poolfile, struct memfile *treefile, const char *fname, int basezoom, int layer, double droprate, long long *file_bbox, json_object *tippecanoe, int segment, int *initialized, unsigned *initial_x, unsigned *initial_y, struct reader *readers, int maxzoom, json_object *feature, std::map<std::string, layermap_entry> *layermap, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types, double *dist_sum, size_t *dist_count, bool want_dist) {
json_object *geometry_type = json_hash_get(geometry, "type");
if (geometry_type == NULL) {
static int warned = 0;
@@ -425,6 +426,33 @@ int serialize_geometry(json_object *geometry, json_object *properties, json_obje
dv = fix_polygon(dv);
}
if (want_dist) {
std::vector<unsigned long long> locs;
for (size_t i = 0; i < dv.size(); i++) {
if (dv[i].op == VT_MOVETO || dv[i].op == VT_LINETO) {
locs.push_back(encode(dv[i].x << geometry_scale, dv[i].y << geometry_scale));
}
}
std::sort(locs.begin(), locs.end());
size_t n = 0;
double sum = 0;
for (size_t i = 1; i < locs.size(); i++) {
if (locs[i - 1] != locs[i]) {
sum += log(locs[i] - locs[i - 1]);
n++;
}
}
if (n > 0) {
double avg = exp(sum / n);
// Convert approximately from tile units to feet
double dist_ft = sqrt(avg) / 33;
*dist_sum += log(dist_ft) * n;
*dist_count += n;
}
locs.clear();
}
bool inline_meta = true;
// Don't inline metadata for features that will span several tiles at maxzoom
if (g > 0 && (bbox[2] < bbox[0] || bbox[3] < bbox[1])) {
@@ -575,7 +603,7 @@ void check_crs(json_object *j, const char *reading) {
}
}
void parse_json(json_pull *jp, const char *reading, volatile long long *layer_seq, volatile long long *progress_seq, long long *metapos, long long *geompos, long long *indexpos, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, FILE *metafile, FILE *geomfile, FILE *indexfile, struct memfile *poolfile, struct memfile *treefile, char *fname, int basezoom, int layer, double droprate, long long *file_bbox, int segment, int *initialized, unsigned *initial_x, unsigned *initial_y, struct reader *readers, int maxzoom, std::map<std::string, layermap_entry> *layermap, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types) {
void parse_json(json_pull *jp, const char *reading, volatile long long *layer_seq, volatile long long *progress_seq, long long *metapos, long long *geompos, long long *indexpos, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, FILE *metafile, FILE *geomfile, FILE *indexfile, struct memfile *poolfile, struct memfile *treefile, char *fname, int basezoom, int layer, double droprate, long long *file_bbox, int segment, int *initialized, unsigned *initial_x, unsigned *initial_y, struct reader *readers, int maxzoom, std::map<std::string, layermap_entry> *layermap, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types, double *dist_sum, size_t *dist_count, bool want_dist) {
long long found_hashes = 0;
long long found_features = 0;
long long found_geometries = 0;
@@ -643,7 +671,7 @@ void parse_json(json_pull *jp, const char *reading, volatile long long *layer_se
}
found_geometries++;
serialize_geometry(j, NULL, NULL, reading, jp->line, layer_seq, progress_seq, metapos, geompos, indexpos, exclude, include, exclude_all, metafile, geomfile, indexfile, poolfile, treefile, fname, basezoom, layer, droprate, file_bbox, NULL, segment, initialized, initial_x, initial_y, readers, maxzoom, j, layermap, layername, uses_gamma, attribute_types);
serialize_geometry(j, NULL, NULL, reading, jp->line, layer_seq, progress_seq, metapos, geompos, indexpos, exclude, include, exclude_all, metafile, geomfile, indexfile, poolfile, treefile, fname, basezoom, layer, droprate, file_bbox, NULL, segment, initialized, initial_x, initial_y, readers, maxzoom, j, layermap, layername, uses_gamma, attribute_types, dist_sum, dist_count, want_dist);
json_free(j);
continue;
}
@@ -686,10 +714,10 @@ void parse_json(json_pull *jp, const char *reading, volatile long long *layer_se
if (geometries != NULL) {
size_t g;
for (g = 0; g < geometries->length; g++) {
serialize_geometry(geometries->array[g], properties, id, reading, jp->line, layer_seq, progress_seq, metapos, geompos, indexpos, exclude, include, exclude_all, metafile, geomfile, indexfile, poolfile, treefile, fname, basezoom, layer, droprate, file_bbox, tippecanoe, segment, initialized, initial_x, initial_y, readers, maxzoom, j, layermap, layername, uses_gamma, attribute_types);
serialize_geometry(geometries->array[g], properties, id, reading, jp->line, layer_seq, progress_seq, metapos, geompos, indexpos, exclude, include, exclude_all, metafile, geomfile, indexfile, poolfile, treefile, fname, basezoom, layer, droprate, file_bbox, tippecanoe, segment, initialized, initial_x, initial_y, readers, maxzoom, j, layermap, layername, uses_gamma, attribute_types, dist_sum, dist_count, want_dist);
}
} else {
serialize_geometry(geometry, properties, id, reading, jp->line, layer_seq, progress_seq, metapos, geompos, indexpos, exclude, include, exclude_all, metafile, geomfile, indexfile, poolfile, treefile, fname, basezoom, layer, droprate, file_bbox, tippecanoe, segment, initialized, initial_x, initial_y, readers, maxzoom, j, layermap, layername, uses_gamma, attribute_types);
serialize_geometry(geometry, properties, id, reading, jp->line, layer_seq, progress_seq, metapos, geompos, indexpos, exclude, include, exclude_all, metafile, geomfile, indexfile, poolfile, treefile, fname, basezoom, layer, droprate, file_bbox, tippecanoe, segment, initialized, initial_x, initial_y, readers, maxzoom, j, layermap, layername, uses_gamma, attribute_types, dist_sum, dist_count, want_dist);
}
json_free(j);
@@ -701,7 +729,7 @@ void parse_json(json_pull *jp, const char *reading, volatile long long *layer_se
void *run_parse_json(void *v) {
struct parse_json_args *pja = (struct parse_json_args *) v;
parse_json(pja->jp, pja->reading, pja->layer_seq, pja->progress_seq, pja->metapos, pja->geompos, pja->indexpos, pja->exclude, pja->include, pja->exclude_all, pja->metafile, pja->geomfile, pja->indexfile, pja->poolfile, pja->treefile, pja->fname, pja->basezoom, pja->layer, pja->droprate, pja->file_bbox, pja->segment, pja->initialized, pja->initial_x, pja->initial_y, pja->readers, pja->maxzoom, pja->layermap, *pja->layername, pja->uses_gamma, pja->attribute_types);
parse_json(pja->jp, pja->reading, pja->layer_seq, pja->progress_seq, pja->metapos, pja->geompos, pja->indexpos, pja->exclude, pja->include, pja->exclude_all, pja->metafile, pja->geomfile, pja->indexfile, pja->poolfile, pja->treefile, pja->fname, pja->basezoom, pja->layer, pja->droprate, pja->file_bbox, pja->segment, pja->initialized, pja->initial_x, pja->initial_y, pja->readers, pja->maxzoom, pja->layermap, *pja->layername, pja->uses_gamma, pja->attribute_types, pja->dist_sum, pja->dist_count, pja->want_dist);
return NULL;
}
+4 -1
View File
@@ -29,10 +29,13 @@ struct parse_json_args {
std::string *layername;
bool uses_gamma;
std::map<std::string, int> const *attribute_types;
double *dist_sum;
size_t *dist_count;
bool want_dist;
};
struct json_pull *json_begin_map(char *map, long long len);
void json_end_map(struct json_pull *jp);
void parse_json(json_pull *jp, const char *reading, volatile long long *layer_seq, volatile long long *progress_seq, long long *metapos, long long *geompos, long long *indexpos, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, FILE *metafile, FILE *geomfile, FILE *indexfile, struct memfile *poolfile, struct memfile *treefile, char *fname, int basezoom, int layer, double droprate, long long *file_bbox, int segment, int *initialized, unsigned *initial_x, unsigned *initial_y, struct reader *readers, int maxzoom, std::map<std::string, layermap_entry> *layermap, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types);
void parse_json(json_pull *jp, const char *reading, volatile long long *layer_seq, volatile long long *progress_seq, long long *metapos, long long *geompos, long long *indexpos, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, FILE *metafile, FILE *geomfile, FILE *indexfile, struct memfile *poolfile, struct memfile *treefile, char *fname, int basezoom, int layer, double droprate, long long *file_bbox, int segment, int *initialized, unsigned *initial_x, unsigned *initial_y, struct reader *readers, int maxzoom, std::map<std::string, layermap_entry> *layermap, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types, double *dist_sum, size_t *dist_count, bool want_dist);
void *run_parse_json(void *v);
+51 -11
View File
@@ -372,7 +372,7 @@ void *run_sort(void *v) {
return NULL;
}
void do_read_parallel(char *map, long long len, long long initial_offset, const char *reading, struct reader *reader, volatile long long *progress_seq, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, char *fname, int basezoom, int source, int nlayers, std::vector<std::map<std::string, layermap_entry> > *layermaps, double droprate, int *initialized, unsigned *initial_x, unsigned *initial_y, int maxzoom, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types, int separator) {
void do_read_parallel(char *map, long long len, long long initial_offset, const char *reading, struct reader *reader, volatile long long *progress_seq, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, char *fname, int basezoom, int source, int nlayers, std::vector<std::map<std::string, layermap_entry> > *layermaps, double droprate, int *initialized, unsigned *initial_x, unsigned *initial_y, int maxzoom, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types, int separator, double *dist_sum, size_t *dist_count, bool want_dist) {
long long segs[CPUS + 1];
segs[0] = 0;
segs[CPUS] = len;
@@ -385,11 +385,15 @@ void do_read_parallel(char *map, long long len, long long initial_offset, const
}
}
double dist_sums[CPUS];
size_t dist_counts[CPUS];
volatile long long layer_seq[CPUS];
for (size_t i = 0; i < CPUS; i++) {
// To preserve feature ordering, unique id for each segment
// begins with that segment's offset into the input
layer_seq[i] = segs[i] + initial_offset;
dist_sums[i] = dist_counts[i] = 0;
}
struct parse_json_args pja[CPUS];
@@ -431,6 +435,9 @@ void do_read_parallel(char *map, long long len, long long initial_offset, const
pja[i].layername = &layername;
pja[i].uses_gamma = uses_gamma;
pja[i].attribute_types = attribute_types;
pja[i].dist_sum = &(dist_sums[i]);
pja[i].dist_count = &(dist_counts[i]);
pja[i].want_dist = want_dist;
if (pthread_create(&pthreads[i], NULL, run_parse_json, &pja[i]) != 0) {
perror("pthread_create");
@@ -445,6 +452,9 @@ void do_read_parallel(char *map, long long len, long long initial_offset, const
perror("pthread_join 370");
}
*dist_sum += dist_sums[i];
*dist_count += dist_counts[i];
json_end_map(pja[i].jp);
}
}
@@ -476,6 +486,9 @@ struct read_parallel_arg {
std::string layername;
bool uses_gamma;
std::map<std::string, int> const *attribute_types;
double *dist_sum;
size_t *dist_count;
bool want_dist;
};
void *run_read_parallel(void *v) {
@@ -497,7 +510,7 @@ void *run_read_parallel(void *v) {
}
madvise(map, rpa->len, MADV_RANDOM); // sequential, but from several pointers at once
do_read_parallel(map, rpa->len, rpa->offset, rpa->reading, rpa->reader, rpa->progress_seq, rpa->exclude, rpa->include, rpa->exclude_all, rpa->fname, rpa->basezoom, rpa->source, rpa->nlayers, rpa->layermaps, rpa->droprate, rpa->initialized, rpa->initial_x, rpa->initial_y, rpa->maxzoom, rpa->layername, rpa->uses_gamma, rpa->attribute_types, rpa->separator);
do_read_parallel(map, rpa->len, rpa->offset, rpa->reading, rpa->reader, rpa->progress_seq, rpa->exclude, rpa->include, rpa->exclude_all, rpa->fname, rpa->basezoom, rpa->source, rpa->nlayers, rpa->layermaps, rpa->droprate, rpa->initialized, rpa->initial_x, rpa->initial_y, rpa->maxzoom, rpa->layername, rpa->uses_gamma, rpa->attribute_types, rpa->separator, rpa->dist_sum, rpa->dist_count, rpa->want_dist);
madvise(map, rpa->len, MADV_DONTNEED);
if (munmap(map, rpa->len) != 0) {
@@ -514,7 +527,7 @@ void *run_read_parallel(void *v) {
return NULL;
}
void start_parsing(int fd, FILE *fp, long long offset, long long len, volatile int *is_parsing, pthread_t *parallel_parser, bool &parser_created, const char *reading, struct reader *reader, volatile long long *progress_seq, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, char *fname, int basezoom, int source, int nlayers, std::vector<std::map<std::string, layermap_entry> > &layermaps, double droprate, int *initialized, unsigned *initial_x, unsigned *initial_y, int maxzoom, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types, int separator) {
void start_parsing(int fd, FILE *fp, long long offset, long long len, volatile int *is_parsing, pthread_t *parallel_parser, bool &parser_created, const char *reading, struct reader *reader, volatile long long *progress_seq, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, char *fname, int basezoom, int source, int nlayers, std::vector<std::map<std::string, layermap_entry> > &layermaps, double droprate, int *initialized, unsigned *initial_x, unsigned *initial_y, int maxzoom, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types, int separator, double *dist_sum, size_t *dist_count, bool want_dist) {
// This has to kick off an intermediate thread to start the parser threads,
// so the main thread can get back to reading the next input stage while
// the intermediate thread waits for the completion of the parser threads.
@@ -553,6 +566,9 @@ void start_parsing(int fd, FILE *fp, long long offset, long long len, volatile i
rpa->layername = layername;
rpa->uses_gamma = uses_gamma;
rpa->attribute_types = attribute_types;
rpa->dist_sum = dist_sum;
rpa->dist_count = dist_count;
rpa->want_dist = want_dist;
if (pthread_create(parallel_parser, NULL, run_read_parallel, rpa) != 0) {
perror("pthread_create");
@@ -1173,6 +1189,8 @@ int read_input(std::vector<source> &sources, char *fname, int &maxzoom, int minz
}
long overall_offset = 0;
double dist_sum = 0;
size_t dist_count = 0;
size_t nsources = sources.size();
for (size_t source = 0; source < nsources; source++) {
@@ -1235,7 +1253,7 @@ int read_input(std::vector<source> &sources, char *fname, int &maxzoom, int minz
}
if (map != NULL && map != MAP_FAILED && read_parallel_this) {
do_read_parallel(map, st.st_size - off, overall_offset, reading.c_str(), reader, &progress_seq, exclude, include, exclude_all, fname, basezoom, layer, nlayers, &layermaps, droprate, initialized, initial_x, initial_y, maxzoom, sources[layer].layer, uses_gamma, attribute_types, read_parallel_this);
do_read_parallel(map, st.st_size - off, overall_offset, reading.c_str(), reader, &progress_seq, exclude, include, exclude_all, fname, basezoom, layer, nlayers, &layermaps, droprate, initialized, initial_x, initial_y, maxzoom, sources[layer].layer, uses_gamma, attribute_types, read_parallel_this, &dist_sum, &dist_count, guess_maxzoom);
overall_offset += st.st_size - off;
checkdisk(reader, CPUS);
@@ -1311,7 +1329,7 @@ int read_input(std::vector<source> &sources, char *fname, int &maxzoom, int minz
}
fflush(readfp);
start_parsing(readfd, readfp, initial_offset, ahead, &is_parsing, &parallel_parser, parser_created, reading.c_str(), reader, &progress_seq, exclude, include, exclude_all, fname, basezoom, layer, nlayers, layermaps, droprate, initialized, initial_x, initial_y, maxzoom, sources[layer].layer, gamma != 0, attribute_types, read_parallel_this);
start_parsing(readfd, readfp, initial_offset, ahead, &is_parsing, &parallel_parser, parser_created, reading.c_str(), reader, &progress_seq, exclude, include, exclude_all, fname, basezoom, layer, nlayers, layermaps, droprate, initialized, initial_x, initial_y, maxzoom, sources[layer].layer, gamma != 0, attribute_types, read_parallel_this, &dist_sum, &dist_count, guess_maxzoom);
initial_offset += ahead;
overall_offset += ahead;
@@ -1348,7 +1366,7 @@ int read_input(std::vector<source> &sources, char *fname, int &maxzoom, int minz
fflush(readfp);
if (ahead > 0) {
start_parsing(readfd, readfp, initial_offset, ahead, &is_parsing, &parallel_parser, parser_created, reading.c_str(), reader, &progress_seq, exclude, include, exclude_all, fname, basezoom, layer, nlayers, layermaps, droprate, initialized, initial_x, initial_y, maxzoom, sources[layer].layer, gamma != 0, attribute_types, read_parallel_this);
start_parsing(readfd, readfp, initial_offset, ahead, &is_parsing, &parallel_parser, parser_created, reading.c_str(), reader, &progress_seq, exclude, include, exclude_all, fname, basezoom, layer, nlayers, layermaps, droprate, initialized, initial_x, initial_y, maxzoom, sources[layer].layer, gamma != 0, attribute_types, read_parallel_this, &dist_sum, &dist_count, guess_maxzoom);
if (parser_created) {
if (pthread_join(parallel_parser, NULL) != 0) {
@@ -1365,7 +1383,7 @@ int read_input(std::vector<source> &sources, char *fname, int &maxzoom, int minz
long long layer_seq = overall_offset;
json_pull *jp = json_begin_file(fp);
parse_json(jp, reading.c_str(), &layer_seq, &progress_seq, &reader[0].metapos, &reader[0].geompos, &reader[0].indexpos, exclude, include, exclude_all, reader[0].metafile, reader[0].geomfile, reader[0].indexfile, reader[0].poolfile, reader[0].treefile, fname, basezoom, layer, droprate, reader[0].file_bbox, 0, &initialized[0], &initial_x[0], &initial_y[0], reader, maxzoom, &layermaps[0], sources[layer].layer, uses_gamma, attribute_types);
parse_json(jp, reading.c_str(), &layer_seq, &progress_seq, &reader[0].metapos, &reader[0].geompos, &reader[0].indexpos, exclude, include, exclude_all, reader[0].metafile, reader[0].geomfile, reader[0].indexfile, reader[0].poolfile, reader[0].treefile, fname, basezoom, layer, droprate, reader[0].file_bbox, 0, &initialized[0], &initial_x[0], &initial_y[0], reader, maxzoom, &layermaps[0], sources[layer].layer, uses_gamma, attribute_types, &dist_sum, &dist_count, guess_maxzoom);
json_end(jp);
overall_offset = layer_seq;
checkdisk(reader, CPUS);
@@ -1618,6 +1636,11 @@ int read_input(std::vector<source> &sources, char *fname, int &maxzoom, int minz
}
}
if (count == 0 && dist_count == 0) {
fprintf(stderr, "Can't guess maxzoom (-zg) without at least two distinct feature locations\n");
exit(EXIT_FAILURE);
}
if (count > 0) {
// Geometric mean is appropriate because distances between features
// are typically lognormally distributed
@@ -1625,7 +1648,8 @@ int read_input(std::vector<source> &sources, char *fname, int &maxzoom, int minz
// Convert approximately from tile units to feet
double dist_ft = sqrt(avg) / 33;
double want = dist_ft / 250;
// Factor of 8 (3 zooms) beyond minimum required to distinguish features
double want = dist_ft / 8;
maxzoom = ceil(log(360 / (.00000274 * want)) / log(2) - full_detail);
if (maxzoom < 0) {
@@ -1638,9 +1662,25 @@ int read_input(std::vector<source> &sources, char *fname, int &maxzoom, int minz
if (!quiet) {
fprintf(stderr, "Choosing a maxzoom of -z%d for features about %d feet apart\n", maxzoom, (int) ceil(dist_ft));
}
} else {
fprintf(stderr, "Can't guess maxzoom (-zg) without at least two distinct feature locations\n");
exit(EXIT_FAILURE);
}
if (dist_count != 0) {
double want2 = exp(dist_sum / dist_count) / 8;
int mz = ceil(log(360 / (.00000274 * want2)) / log(2) - full_detail);
if (mz < 0) {
mz = 0;
}
if (mz > MAX_ZOOM) {
mz = MAX_ZOOM;
}
if (mz > maxzoom || count <= 0) {
if (!quiet) {
fprintf(stderr, "Choosing a maxzoom of -z%d for resolution of about %d feet within features\n", mz, (int) exp(dist_sum / dist_count));
}
maxzoom = mz;
}
}
if (maxzoom < minzoom) {
File diff suppressed because one or more lines are too long
File diff suppressed because it is too large Load Diff
File diff suppressed because it is too large Load Diff
File diff suppressed because one or more lines are too long
+1 -1
View File
@@ -1834,7 +1834,7 @@ long long write_tile(FILE *geoms, long long *geompos_in, char *metabase, char *s
line_detail++; // to keep it the same when the loop decrements it
continue;
} else {
fprintf(stderr, "Try using -B (and --drop-lines or --drop-polygons if needed) to set a higher base zoom level.\n");
fprintf(stderr, "Try using --drop-fraction-as-needed or --drop-densest-as-needed.\n");
return -1;
}
}
+1 -1
View File
@@ -1 +1 @@
#define VERSION "tippecanoe v1.18.2\n"
#define VERSION "tippecanoe v1.19.0\n"