Add --reject-null-geometry and more GeoJSON geometry validation

CHANGELOG.md

@@ -1,3 +1,7 @@
+# 2.38.0
+
+* Add --reject-null-geometry and more checking of GeoJSON polygon and linestring validity
+
 # 2.37.0
 
 * Speed up tile-join overzooming and make it use less memory, by not including empty child tiles in the enumeration

README.md

@@ -533,6 +533,7 @@ the same layer, enclose them in an `all` expression so they will all be evaluate
  * `-pW` or `--reverse-source-polygon-winding`: Instead of respecting GeoJSON polygon ring order, use the opposite of the original polygon winding in the source data to distinguish inner (counterclockwise) and outer (clockwise) polygon rings.
  * `--clip-bounding-box=`*minlon*`,`*minlat*`,`*maxlon*`,`*maxlat*: Clip all features to the specified bounding box.
  * `-aP` or `--convert-polygons-to-label-points`: Replace polygon geometries with a label point or points for the polygon in each tile it intersects.
+ * `--reject-null-geometry`: Exit with an error when a feature has a `null` geometry instead of just giving a warning.
 
 ### Setting or disabling tile size limits
 

errors.hpp

@@ -19,5 +19,6 @@
 #define EXIT_UNLINK 118
 #define EXIT_UTF8 119
 #define EXIT_WRITE 120
+#define EXIT_INCORRECT_GEOMETRY 121
 
 // avoid 124, 125, 126, 127, 137, which are used by GNU timeout

geojson.cpp

@@ -49,6 +49,10 @@ int serialize_geojson_feature(struct serialization_state *sst, json_object *geom
 			fprintf(stderr, "%s:%d: null geometry (additional not reported): ", sst->fname, sst->line);
 			json_context(feature);
 			warned = 1;
+
+			if (additional[A_REJECT_NULL_GEOMETRY]) {
+				exit(EXIT_INCORRECT_GEOMETRY);
+			}
 		}
 
 		return 0;

main.cpp

@@ -3214,6 +3214,7 @@ int main(int argc, char **argv) {
 		{"reverse-source-polygon-winding", no_argument, &prevent[P_REVERSE_SOURCE_POLYGON_WINDING], 1},
 		{"clip-bounding-box", required_argument, 0, '~'},
 		{"convert-polygons-to-label-points", no_argument, &additional[A_GENERATE_POLYGON_LABEL_POINTS], 1},
+		{"reject-null-geometry", no_argument, &additional[A_REJECT_NULL_GEOMETRY], 1},
 
 		{"Filtering tile contents", 0, 0, 0},
 		{"prefilter", required_argument, 0, 'C'},

man/tippecanoe.1

@@ -695,6 +695,8 @@ the line or polygon within one tile unit of its proper location. You can probabl
 \fB\fC\-\-clip\-bounding\-box=\fR\fIminlon\fP\fB\fC,\fR\fIminlat\fP\fB\fC,\fR\fImaxlon\fP\fB\fC,\fR\fImaxlat\fP: Clip all features to the specified bounding box.
 .IP \(bu 2
 \fB\fC\-aP\fR or \fB\fC\-\-convert\-polygons\-to\-label\-points\fR: Replace polygon geometries with a label point or points for the polygon in each tile it intersects.
+.IP \(bu 2
+\fB\fC\-\-reject\-null\-geometry\fR: Exit with an error when a feature has a \fB\fCnull\fR geometry instead of just giving a warning.
 .RE
 .SS Setting or disabling tile size limits
 .RS

options.hpp

@@ -26,6 +26,7 @@
 #define A_HILBERT ((int) 'h')
 #define A_VISVALINGAM ((int) 'v')
 #define A_GENERATE_POLYGON_LABEL_POINTS ((int) 'P')
+#define A_REJECT_NULL_GEOMETRY ((int) '0')
 
 #define P_SIMPLIFY ((int) 's')
 #define P_SIMPLIFY_LOW ((int) 'S')

read_json.cpp

@@ -103,6 +103,14 @@ void parse_geometry(int t, json_object *j, drawvec &out, int op, const char *fna
 	}
 
 	if (t == GEOM_POLYGON) {
+		if (out.size() < 3) {
+			fprintf(stderr, "%s:%d: polygon must contain at least three points: ", fname, line);
+			json_context(j);
+			fprintf(stderr, "%s:%d: polygon must contain at least three points: ", fname, line);
+			json_context(feature);
+			exit(EXIT_INCORRECT_GEOMETRY);
+		}
+
 		// Note that this is not using the correct meaning of closepath.
 		//
 		// We are using it here to close an entire Polygon, to distinguish
@@ -114,6 +122,16 @@ void parse_geometry(int t, json_object *j, drawvec &out, int op, const char *fna
 
 		out.push_back(draw(VT_CLOSEPATH, 0, 0));
 	}
+
+	if (t == GEOM_LINESTRING) {
+		if (out.size() < 2) {
+			fprintf(stderr, "%s:%d: linestring must contain at least two points: ", fname, line);
+			json_context(j);
+			fprintf(stderr, "%s:%d: linestring must contain at least two points: ", fname, line);
+			json_context(feature);
+			exit(EXIT_INCORRECT_GEOMETRY);
+		}
+	}
 }
 
 void stringify_value(json_object *value, int &type, std::string &stringified, const char *reading, int line, json_object *feature) {

tile.cpp

@@ -1609,6 +1609,34 @@ serial_feature next_feature(decompressor *geoms, std::atomic<long long> *geompos
 	}
 }
 
+bool legit_geometry(int t, const drawvec &g) {
+	bool legit = false;
+
+	for (size_t i = 0; i < g.size(); i++) {
+		if (g[i].op == VT_MOVETO) {
+			size_t j;
+
+			for (j = i + 1; j < g.size(); j++) {
+				if (g[j].op != VT_LINETO) {
+					break;
+				}
+			}
+
+			if (t == VT_LINE && j < i + 2) {
+				return false;
+			}
+			if (t == VT_POLYGON && j < i + 3) {
+				return false;
+			}
+
+			legit = true;  // must contain at least one ring
+			i = j - 1;
+		}
+	}
+
+	return legit;
+}
+
 struct run_prefilter_args {
 	decompressor *geoms = NULL;
 	std::atomic<long long> *geompos_in = NULL;
@@ -1657,6 +1685,11 @@ void *run_prefilter(void *v) {
 		tmp_layer.extent = 1LL << 32;
 		tmp_layer.name = (*(rpa->layer_unmaps))[sf.segment][sf.layer];
 
+		// make sure the geometry is actually legal geojson
+		if (!legit_geometry(sf.t, sf.geometry)) {
+			continue;
+		}
+
 		if (sf.t == VT_POLYGON) {
 			sf.geometry = close_poly(sf.geometry);
 		}

version.hpp

@@ -1,6 +1,6 @@
 #ifndef VERSION_HPP
 #define VERSION_HPP
 
-#define VERSION "v2.37.0"
+#define VERSION "v2.38.0"
 
 #endif