feat: enable usage of csv files with wkt geometries

Makefile

@@ -549,6 +549,11 @@ csv-test: tippecanoe tippecanoe-decode
 	./tippecanoe-decode -x generator -x generator_options tests/csv/out.mbtiles > tests/csv/out.mbtiles.json.check
 	cmp tests/csv/out.mbtiles.json.check tests/csv/out.mbtiles.json
 	rm -f tests/csv/out.mbtiles.json.check tests/csv/out.mbtiles
+	# Reading from named CSV with geometry
+	./tippecanoe -q -zg -f -o tests/csv_geometry/out.mbtiles tests/csv_geometry/ne_110m_populated_places_simple.csv
+	./tippecanoe-decode -x generator -x generator_options tests/csv_geometry/out.mbtiles > tests/csv_geometry/out.mbtiles.json.check
+	cmp tests/csv_geometry/out.mbtiles.json.check tests/csv_geometry/out.mbtiles.json
+	rm -f tests/csv_geometry/out.mbtiles.json.check tests/csv_geometry/out.mbtiles
 
 layer-json-test: tippecanoe tippecanoe-decode
 	# GeoJSON with description and named layer

README.md

@@ -4,7 +4,7 @@ tippecanoe
 Builds [vector tilesets](https://github.com/mapbox/vector-tile-spec/) from large (or small) collections of [GeoJSON](http://geojson.org/), [FlatGeobuf](https://github.com/flatgeobuf/flatgeobuf), or [CSV](https://en.wikipedia.org/wiki/Comma-separated_values) features,
 [like these](MADE_WITH.md).
 
-This is the official home of Tippecanoe, developed and actively maintained by [Erica Fischer](https://github.com/e-n-f) at [Felt](https://felt.com). 
+This is the official home of Tippecanoe, developed and actively maintained by [Erica Fischer](https://github.com/e-n-f) at [Felt](https://felt.com).
 
 Version 2.0.0 is equivalent to [1.36.0](https://github.com/mapbox/tippecanoe/tree/1.36.0) in the original repository. Thank you Mapbox for the many years of early support.
 
@@ -317,7 +317,7 @@ If your input is formatted as newline-delimited GeoJSON, use `-P` to make input
 tippecanoe -z5 -o world.mbtiles -L'{"file":"ne_10m_admin_0_countries.json", "layer":"countries", "description":"Natural Earth countries"}'
 ```
 
-CSV input files currently support only Point geometries, from columns named `latitude`, `longitude`, `lat`, `lon`, `long`, `lng`, `x`, or `y`.
+CSV input files currently support Point geometries, from columns named `latitude`, `longitude`, `lat`, `lon`, `long`, `lng`, `x`, or `y`and Point, LineString, and Polygon geometries as a WKT string from columns named `geometry` or `wkt`.
 
 ### Parallel processing of input
 

geocsv.cpp

@@ -1,3 +1,7 @@
+#include <sstream>
+#include <vector>
+#include <string>
+#include <cctype>
 #include <stdlib.h>
 #include <algorithm>
 #include "geocsv.hpp"
@@ -11,6 +15,103 @@
 #include "options.hpp"
 #include "errors.hpp"
 
+#include <iostream>
+#include <vector>
+#include <string>
+#include <sstream>
+#include <cctype>
+
+void trim(std::string &s) {
+	s.erase(0, s.find_first_not_of(' '));
+	s.erase(s.find_last_not_of(' ') + 1);
+}
+
+void processRing(const std::string &ring, drawvec &dv) {
+	std::stringstream coordStream(ring);
+	std::string coord;
+	bool first = true;
+	while (std::getline(coordStream, coord, ',')) {
+		coord.erase(0, coord.find_first_not_of(' '));
+		coord.erase(coord.find_last_not_of(' ') + 1);
+		std::stringstream pointStream(coord);
+		std::string xStr, yStr;
+		std::getline(pointStream, xStr, ' ');
+		std::getline(pointStream, yStr);
+		long long wx, wy;
+		double x = std::stod(xStr);
+		double y = std::stod(yStr);
+		projection->project(x, y, 32, &wx, &wy);
+		if (first) {
+			dv.push_back(draw(VT_MOVETO, wx, wy));
+			first = false;
+		} else {
+			dv.push_back(draw(VT_LINETO, wx, wy));
+		}
+	}
+	dv.push_back(draw(VT_CLOSEPATH, 0, 0));
+}
+
+drawvec parse_wkt(const std::string &wkt, drawvec &dv, int &geometry_type) {
+	std::string type, coordinates;
+	std::stringstream ss(wkt);
+
+	// Read geometry type
+	std::getline(ss, type, '(');
+	type.erase(0, type.find_first_not_of(' '));
+	type.erase(type.find_last_not_of(' ') + 1);
+
+	// Read coordinates
+	std::getline(ss, coordinates);
+	coordinates = coordinates.substr(0, coordinates.size() - 1);
+	coordinates.erase(0, coordinates.find_first_not_of(' '));
+	coordinates.erase(coordinates.find_last_not_of(' ') + 1);
+
+	std::stringstream coordStream(coordinates);
+
+	if (type == "POINT") {
+		geometry_type = VT_POINT;
+		std::string xStr, yStr;
+		std::getline(coordStream, xStr, ' ');
+		std::getline(coordStream, yStr);
+		trim(xStr);
+		trim(yStr);
+		long long wx, wy;
+		double x = std::stod(xStr);
+		double y = std::stod(yStr);
+		projection->project(x, y, 32, &wx, &wy);
+		dv.push_back(draw(VT_MOVETO, wx, wy));
+	} else if (type == "LINESTRING") {
+		geometry_type = VT_LINE;
+		processRing(coordinates, dv);
+	} else if (type == "POLYGON") {
+		geometry_type = VT_POLYGON;
+		// Handle POLYGON type with multiple rings
+		std::vector<std::string> rings;
+		std::string ring;
+		int level = 0;
+		for (char c : coordinates) {
+			if (c == '(') {
+				if (level == 0) ring.clear();
+				level++;
+			} else if (c == ')') {
+				level--;
+				if (level == 0) {
+					trim(ring);
+					rings.push_back(ring);
+				}
+			}
+			if (level > 0 && c != '(') {
+				ring += c;
+			}
+		}
+		for (const std::string &currentRing : rings) {
+			processRing(currentRing, dv);
+		}
+	}
+
+	return dv;
+}
+
 void parse_geocsv(std::vector<struct serialization_state> &sst, std::string fname, int layer, std::string layername) {
 	FILE *f;
 
@@ -26,7 +127,7 @@ void parse_geocsv(std::vector<struct serialization_state> &sst, std::string fnam
 
 	std::string s;
 	std::vector<std::string> header;
-	ssize_t latcol = -1, loncol = -1;
+	ssize_t latcol = -1, loncol = -1, geometrycol = -1;
 
 	if ((s = csv_getline(f)).size() > 0) {
 		std::string err = check_utf8(s);
@@ -49,11 +150,14 @@ void parse_geocsv(std::vector<struct serialization_state> &sst, std::string fnam
 			if (lower == "x" || lower == "lon" || lower == "lng" || lower == "long" || (lower.find("longitude") != std::string::npos)) {
 				loncol = i;
 			}
+			if (lower == "geometry" || lower == "wkt") {
+				geometrycol = i;
+			}
 		}
 	}
 
-	if (latcol < 0 || loncol < 0) {
-		fprintf(stderr, "%s: Can't find \"lat\" and \"lon\" columns\n", fname.c_str());
+	if ((latcol < 0 || loncol < 0) && geometrycol < 0) {
+		fprintf(stderr, "%s: Can't find \"lat\" and \"lon\" or \"geometry\" columns\n", fname.c_str());
 		exit(EXIT_CSV);
 	}
 
@@ -73,27 +177,33 @@ void parse_geocsv(std::vector<struct serialization_state> &sst, std::string fnam
 			exit(EXIT_CSV);
 		}
 
-		if (line[loncol].empty() || line[latcol].empty()) {
+		if ((line[loncol].empty() || line[latcol].empty()) && line[geometrycol].empty()) {
 			static int warned = 0;
 			if (!warned) {
 				fprintf(stderr, "%s:%zu: null geometry (additional not reported)\n", fname.c_str(), seq + 1);
 				warned = 1;
 			}
 			continue;
 		}
-		double lon = atof(line[loncol].c_str());
-		double lat = atof(line[latcol].c_str());
-
-		long long x, y;
-		projection->project(lon, lat, 32, &x, &y);
 		drawvec dv;
-		dv.push_back(draw(VT_MOVETO, x, y));
+		int geometry_type = -1;
+		if (latcol >= 0 && loncol >= 0) {
+			double lon = atof(line[loncol].c_str());
+			double lat = atof(line[latcol].c_str());
+
+			long long x, y;
+			projection->project(lon, lat, 32, &x, &y);
+			dv.push_back(draw(VT_MOVETO, x, y));
+			geometry_type = VT_POINT;
+		} else if (geometrycol >= 0) {
+			parse_wkt(csv_dequote(line[geometrycol]), dv, geometry_type);
+		}
 
 		std::vector<std::string> full_keys;
 		std::vector<serial_val> full_values;
 
 		for (size_t i = 0; i < line.size(); i++) {
-			if (i != (size_t) latcol && i != (size_t) loncol) {
+			if (i != (size_t) latcol && i != (size_t) loncol && i != (size_t) geometrycol) {
 				line[i] = csv_dequote(line[i]);
 
 				serial_val sv;
@@ -123,7 +233,7 @@ void parse_geocsv(std::vector<struct serialization_state> &sst, std::string fnam
 		sf.feature_minzoom = false;
 		sf.seq = *(sst[0].layer_seq);
 		sf.geometry = dv;
-		sf.t = 1;  // POINT
+		sf.t = geometry_type;
 		sf.full_keys = full_keys;
 		sf.full_values = full_values;
 

man/tippecanoe.1

@@ -3,7 +3,7 @@
 Builds vector tilesets \[la]https://github.com/mapbox/vector-tile-spec/\[ra] from large (or small) collections of GeoJSON \[la]http://geojson.org/\[ra], FlatGeobuf \[la]https://github.com/flatgeobuf/flatgeobuf\[ra], or CSV \[la]https://en.wikipedia.org/wiki/Comma-separated_values\[ra] features,
 like these \[la]MADE_WITH.md\[ra]\&.
 .PP
-This is the official home of Tippecanoe, developed and actively maintained by Erica Fischer \[la]https://github.com/e-n-f\[ra] at Felt \[la]https://felt.com\[ra]\&. 
+This is the official home of Tippecanoe, developed and actively maintained by Erica Fischer \[la]https://github.com/e-n-f\[ra] at Felt \[la]https://felt.com\[ra]\&.
 .PP
 Version 2.0.0 is equivalent to 1.36.0 \[la]https://github.com/mapbox/tippecanoe/tree/1.36.0\[ra] in the original repository. Thank you Mapbox for the many years of early support.
 .SH Intent
@@ -374,7 +374,7 @@ tippecanoe \-z5 \-o world.mbtiles \-L'{"file":"ne_10m_admin_0_countries.json", "
 .fi
 .RE
 .PP
-CSV input files currently support only Point geometries, from columns named \fB\fClatitude\fR, \fB\fClongitude\fR, \fB\fClat\fR, \fB\fClon\fR, \fB\fClong\fR, \fB\fClng\fR, \fB\fCx\fR, or \fB\fCy\fR\&.
+CSV input files currently support Point geometries, from columns named \fB\fClatitude\fR, \fB\fClongitude\fR, \fB\fClat\fR, \fB\fClon\fR, \fB\fClong\fR, \fB\fClng\fR, \fB\fCx\fR, or \fB\fCy\fRand Point, LineString, and Polygon geometries as a WKT string from columns named \fB\fCgeometry\fR or \fB\fCwkt\fR\&.
 .SS Parallel processing of input
 .RS
 .IP \(bu 2
@@ -471,8 +471,8 @@ zoom level	precision (ft)	precision (m)	map scale
 \fB\fC\-z18\fR	1.5 in	4 cm	1:1250
 \fB\fC\-z19\fR	0.8 in	2 cm	1:600
 \fB\fC\-z20\fR	0.4 in	1 cm	1:300
-\fB\fC\-z21\fR	0.2 in	0.5 cm	1:150
-\fB\fC\-z22\fR	0.1 in	0.25 cm	1:75
+\fB\fC\-z21\fR	0.4 in	1 cm	1:300
+\fB\fC\-z22\fR	0.4 in	1 cm	1:300
 .TE
 .SS Tile resolution
 .RS