Merge branch 'master' into add-build-workflows

README.md

@@ -1,12 +1,23 @@
 # osm2rdf
 
-`osm2rdf` is a tool for converting the complete [OpenStreetMap](https://www.openstreetmap.org) (OSM) data to [RDF Turtle](https://www.w3.org/TR/turtle) (TTL). The produced triples include:
+`osm2rdf` is a tool for converting the complete [OpenStreetMap](https://www.openstreetmap.org) (OSM) data to [RDF Turtle](https://www.w3.org/TR/turtle) (TTL). 
 
-1. One triple `<o> <key> <value>` for each key-value pair (called "tag" in OSM) of each object.
-2. One triple `<o> geo:hasGeometry <wkt>` for the shape of each object (using [WKT](https://en.wikipedia.org/wiki/Well-known_text_representation_of_geometry) to represent the shape).
-3. One triple `<o1> geo:sfContains <o2>` for each pair of objects, where `<o1>` geometrically contains `<o2>`.
-4. One triple `<o1> geo:sfContains <o2>` for each pair of objects, where `<o1>` geometrically intersects `<o2>`.
-5. Triples for various other kinds of information (see `osm2rdf --help` for the many options of the tool).
+We consider the following prefixes being used in our code and documentation:
+
+```
+PREFIX ogc: <http://www.opengis.net/rdf#>
+PREFIX geo: <http://www.opengis.net/ont/geosparql#>
+PREFIX osm2rdf: <https://osm2rdf.cs.uni-freiburg.de/rdf#>
+```
+
+The produced triples include:
+
+1. One triple `<o> <key> <value> .` for each key-value pair (called "tag" in OSM) of each object.
+2. Two triples `<o> geo:hasGeometry <geom_o> . <geom_o> geo:asWKT <wkt> .` for the shape of each object (using [WKT](https://en.wikipedia.org/wiki/Well-known_text_representation_of_geometry) to represent the shape).
+   > Note: using the option `--hasgeometry-as-wkt` will attach the WKT literal directly to the object, i.e. `<o> geo:hasGeometry <wkt> .`
+4. One triple `<o1> ogc:sfContains <o2> .` for each pair of objects, where `<o1>` geometrically contains `<o2>`.
+5. One triple `<o1> ogc:sfIntersects <o2> .` for each pair of objects, where `<o1>` geometrically intersects `<o2>`.
+6. Triples for various other kinds of information (see `osm2rdf --help` for the many options of the tool).
 
 For the complete OSM data, the tool takes around 15 hours on standard hardware and produces around 40 billion triples with a size of around 200 GB for the compressed TTL file. 
 

include/osm2rdf/config/Config.h

@@ -66,6 +66,7 @@ struct Config {
   bool addAreaConvexHull = false;
   bool addAreaEnvelope = false;
   bool addAreaOrientedBoundingBox = false;
+  bool addAreaWayLinestrings = false;
   bool addNodeConvexHull = false;
   bool addNodeEnvelope = false;
   bool addNodeOrientedBoundingBox = false;

include/osm2rdf/config/Constants.h

@@ -201,6 +201,14 @@ const static inline std::string ADD_AREA_ORIENTED_BOUNDING_BOX_OPTION_LONG =
 const static inline std::string ADD_AREA_ORIENTED_BOUNDING_BOX_OPTION_HELP =
     "Add oriented-bounding-box to areas";
 
+const static inline std::string ADD_AREA_WAY_LINESTRINGS_INFO =
+    "Adding linestrings for ways which form areas";
+const static inline std::string ADD_AREA_WAY_LINESTRINGS_OPTION_SHORT = "";
+const static inline std::string ADD_AREA_WAY_LINESTRINGS_OPTION_LONG =
+    "add-area-way-linestrings";
+const static inline std::string ADD_AREA_WAY_LINESTRINGS_OPTION_HELP =
+    "Add linestrings for ways which form areas";
+
 const static inline std::string HASGEOMETRY_AS_WKT_INFO =
     "Letting geo:hasGeometry point directly to WKT serialization literal";
 const static inline std::string HASGEOMETRY_AS_WKT_OPTION_SHORT = "";

include/osm2rdf/osm/GeometryHandler.h

@@ -57,7 +57,6 @@ struct GeomRelationStats {
   size_t _skippedByBoxIdIntersectCutout = 0;
   size_t _skippedByContainedInInnerRing = 0;
   size_t _skippedByBorderContained = 0;
-  size_t _skippedByNodeContained = 0;
   size_t _skippedByInner = 0;
   size_t _skippedByOuter = 0;
   size_t _skippedByBox = 0;
@@ -79,7 +78,6 @@ struct GeomRelationStats {
     _skippedByOrientedBox += lh._skippedByOrientedBox;
     _skippedByConvexHull += lh._skippedByConvexHull;
     _skippedByBorderContained += lh._skippedByBorderContained;
-    _skippedByNodeContained += lh._skippedByNodeContained;
     return *this;
   }
 
@@ -96,7 +94,6 @@ struct GeomRelationStats {
   void skippedByOrientedBox() { _skippedByOrientedBox++; }
   void skippedByConvexHull() { _skippedByConvexHull++; }
   void skippedByBorderContained() { _skippedByBorderContained++; }
-  void skippedByNodeContained() { _skippedByNodeContained++; }
   void fullCheck() { _fullChecks++; }
 
   [[nodiscard]] std::string printPercNum(size_t n) const {
@@ -146,9 +143,6 @@ struct GeomRelationStats {
   [[nodiscard]] std::string printSkippedByOuter() const {
     return printPercNum(_skippedByOuter);
   }
-  [[nodiscard]] std::string printSkippedByNodeContained() const {
-    return printPercNum(_skippedByNodeContained);
-  }
   [[nodiscard]] std::string printFullChecks() const {
     return printPercNum(_fullChecks);
   }
@@ -299,16 +293,15 @@ class GeometryHandler {
   FRIEND_TEST(OSM_GeometryHandler, dumpUnnamedAreaRelationsSimpleContainsOnly);
 
   // Calculate relations for each node.
-  NodesContainedInAreasData dumpNodeRelations();
+  void dumpNodeRelations();
   FRIEND_TEST(OSM_GeometryHandler, noNodeGeometricRelations);
   FRIEND_TEST(OSM_GeometryHandler, dumpNodeRelationsEmpty1);
   FRIEND_TEST(OSM_GeometryHandler, dumpNodeRelationsEmpty2);
   FRIEND_TEST(OSM_GeometryHandler, dumpNodeRelationsSimpleIntersects);
   FRIEND_TEST(OSM_GeometryHandler, dumpNodeRelationsSimpleContains);
 
   // Calculate relations for each way.
-  void dumpWayRelations(
-      const osm2rdf::osm::NodesContainedInAreasData& nodeData);
+  void dumpWayRelations();
   FRIEND_TEST(OSM_GeometryHandler, noWayGeometricRelations);
   FRIEND_TEST(OSM_GeometryHandler, dumpWayRelationsEmpty1);
   FRIEND_TEST(OSM_GeometryHandler, dumpWayRelationsEmpty2);

include/osm2rdf/osm/Relation.h

@@ -45,6 +45,7 @@ class Relation {
       const noexcept;
   [[nodiscard]] const osm2rdf::osm::TagList& tags() const noexcept;
   [[nodiscard]] bool hasCompleteGeometry() const noexcept;
+  [[nodiscard]] bool isArea() const noexcept;
 #if BOOST_VERSION >= 107800
   [[nodiscard]] bool hasGeometry() const noexcept;
   [[nodiscard]] const osm2rdf::geometry::Box& envelope() const noexcept;

include/osm2rdf/osm/Way.h

@@ -45,6 +45,7 @@ class Way {
   explicit Way(const osmium::Way& way);
   [[nodiscard]] id_t id() const noexcept;
   [[nodiscard]] bool closed() const noexcept;
+  [[nodiscard]] bool isArea() const noexcept;
   [[nodiscard]] const osm2rdf::geometry::Box& envelope() const noexcept;
   [[nodiscard]] const osm2rdf::geometry::Way& geom() const noexcept;
   // Return the convex hull.

src/config/Config.cpp

@@ -17,13 +17,14 @@
 // You should have received a copy of the GNU General Public License
 // along with osm2rdf.  If not, see <https://www.gnu.org/licenses/>.
 
+#include "osm2rdf/config/Config.h"
+
 #include <filesystem>
 #include <iostream>
 #include <string>
 
 #include "boost/version.hpp"
 #include "omp.h"
-#include "osm2rdf/config/Config.h"
 #include "osm2rdf/config/Constants.h"
 #include "osm2rdf/config/ExitCode.h"
 #include "popl.hpp"
@@ -150,6 +151,11 @@ std::string osm2rdf::config::Config::getInfo(std::string_view prefix) const {
             << prefix
             << osm2rdf::config::constants::ADD_WAY_NODE_SPATIAL_METADATA_INFO;
       }
+      if (addAreaWayLinestrings) {
+        oss << "\n"
+            << prefix
+            << osm2rdf::config::constants::ADD_AREA_WAY_LINESTRINGS_INFO;
+      }
     }
     if (simplifyWKT > 0) {
       oss << "\n" << prefix << osm2rdf::config::constants::SIMPLIFY_WKT_INFO;
@@ -335,6 +341,11 @@ void osm2rdf::config::Config::fromArgs(int argc, char** argv) {
           osm2rdf::config::constants::ADD_AREA_ENVELOPE_RATIO_OPTION_SHORT,
           osm2rdf::config::constants::ADD_AREA_ENVELOPE_RATIO_OPTION_LONG,
           osm2rdf::config::constants::ADD_AREA_ENVELOPE_RATIO_OPTION_HELP);
+  auto addAreaWayLinestringsOp =
+      parser.add<popl::Switch, popl::Attribute::expert>(
+          osm2rdf::config::constants::ADD_AREA_WAY_LINESTRINGS_OPTION_SHORT,
+          osm2rdf::config::constants::ADD_AREA_WAY_LINESTRINGS_OPTION_LONG,
+          osm2rdf::config::constants::ADD_AREA_WAY_LINESTRINGS_OPTION_HELP);
   auto addRelationBorderMembersOp = parser.add<popl::Switch>(
       osm2rdf::config::constants::ADD_RELATION_BORDER_MEMBERS_OPTION_SHORT,
       osm2rdf::config::constants::ADD_RELATION_BORDER_MEMBERS_OPTION_LONG,
@@ -571,6 +582,7 @@ void osm2rdf::config::Config::fromArgs(int argc, char** argv) {
     addAreaEnvelope = addAreaEnvelopeOp->is_set();
     addAreaEnvelopeRatio = addAreaEnvelopeRatioOp->is_set();
     addAreaOrientedBoundingBox = addAreaOrientedBoundingBoxOp->is_set();
+    addAreaWayLinestrings = addAreaWayLinestringsOp->is_set();
     addRelationBorderMembers = addRelationBorderMembersOp->is_set();
 #if BOOST_VERSION >= 107800
     addRelationConvexHull = addRelationConvexHullOp->is_set();

src/osm/FactHandler.cpp

@@ -17,6 +17,8 @@
 // You should have received a copy of the GNU General Public License
 // along with osm2rdf.  If not, see <https://www.gnu.org/licenses/>.
 
+#include "osm2rdf/osm/FactHandler.h"
+
 #include <iomanip>
 #include <iostream>
 
@@ -25,7 +27,6 @@
 #include "osm2rdf/config/Config.h"
 #include "osm2rdf/osm/Area.h"
 #include "osm2rdf/osm/Constants.h"
-#include "osm2rdf/osm/FactHandler.h"
 #include "osm2rdf/osm/Node.h"
 #include "osm2rdf/osm/Relation.h"
 #include "osm2rdf/osm/Way.h"
@@ -36,10 +37,10 @@ using osm2rdf::osm::constants::BASE_SIMPLIFICATION_FACTOR;
 using osm2rdf::ttl::constants::IRI__GEOSPARQL__AS_WKT;
 using osm2rdf::ttl::constants::IRI__GEOSPARQL__HAS_GEOMETRY;
 using osm2rdf::ttl::constants::IRI__GEOSPARQL__WKT_LITERAL;
+using osm2rdf::ttl::constants::IRI__OSM2RDF__POS;
 using osm2rdf::ttl::constants::IRI__OSM2RDF_GEOM__CONVEX_HULL;
 using osm2rdf::ttl::constants::IRI__OSM2RDF_GEOM__ENVELOPE;
 using osm2rdf::ttl::constants::IRI__OSM2RDF_GEOM__OBB;
-using osm2rdf::ttl::constants::IRI__OSM2RDF__POS;
 using osm2rdf::ttl::constants::IRI__OSM_NODE;
 using osm2rdf::ttl::constants::IRI__OSM_RELATION;
 using osm2rdf::ttl::constants::IRI__OSM_TAG;
@@ -81,7 +82,7 @@ void osm2rdf::osm::FactHandler<W>::area(const osm2rdf::osm::Area& area) {
   if (!_config.hasGeometryAsWkt) {
     const std::string& geomObj = _writer->generateIRI(
         NAMESPACE__OSM2RDF_GEOM, (area.fromWay() ? "wayarea_" : "relarea_") +
-                                std::to_string(area.objId()));
+                                     std::to_string(area.objId()));
 
     _writer->writeTriple(subj, IRI__GEOSPARQL__HAS_GEOMETRY, geomObj);
     writeBoostGeometry(geomObj, IRI__GEOSPARQL__AS_WKT, area.geom());
@@ -96,7 +97,8 @@ void osm2rdf::osm::FactHandler<W>::area(const osm2rdf::osm::Area& area) {
     writeBox(subj, IRI__OSM2RDF_GEOM__ENVELOPE, area.envelope());
   }
   if (_config.addAreaOrientedBoundingBox) {
-    writeBoostGeometry(subj, IRI__OSM2RDF_GEOM__OBB, area.orientedBoundingBox());
+    writeBoostGeometry(subj, IRI__OSM2RDF_GEOM__OBB,
+                       area.orientedBoundingBox());
   }
 
   if (_config.addSortMetadata) {
@@ -146,7 +148,8 @@ void osm2rdf::osm::FactHandler<W>::node(const osm2rdf::osm::Node& node) {
     writeBox(subj, IRI__OSM2RDF_GEOM__ENVELOPE, node.envelope());
   }
   if (_config.addNodeOrientedBoundingBox) {
-    writeBoostGeometry(subj, IRI__OSM2RDF_GEOM__OBB, node.orientedBoundingBox());
+    writeBoostGeometry(subj, IRI__OSM2RDF_GEOM__OBB,
+                       node.orientedBoundingBox());
   }
 }
 
@@ -200,7 +203,7 @@ void osm2rdf::osm::FactHandler<W>::relation(
   }
 
 #if BOOST_VERSION >= 107800
-  if (relation.hasGeometry()) {
+  if (relation.hasGeometry() && !relation.isArea()) {
     if (!_config.hasGeometryAsWkt) {
       const std::string& geomObj = _writer->generateIRI(
           NAMESPACE__OSM2RDF_GEOM, "relation_" + std::to_string(relation.id()));
@@ -305,14 +308,16 @@ void osm2rdf::osm::FactHandler<W>::way(const osm2rdf::osm::Way& way) {
   osm2rdf::geometry::Linestring locations{way.geom()};
   size_t numUniquePoints = locations.size();
 
-  if (!_config.hasGeometryAsWkt) {
-    const std::string& geomObj = _writer->generateIRI(
-        NAMESPACE__OSM2RDF, "way_" + std::to_string(way.id()));
+  if (_config.addAreaWayLinestrings || !way.isArea()) {
+    if (!_config.hasGeometryAsWkt) {
+      const std::string& geomObj = _writer->generateIRI(
+          NAMESPACE__OSM2RDF, "way_" + std::to_string(way.id()));
 
-    _writer->writeTriple(subj, IRI__GEOSPARQL__HAS_GEOMETRY, geomObj);
-    writeBoostGeometry(geomObj, IRI__GEOSPARQL__AS_WKT, locations);
-  } else {
-    writeBoostGeometry(subj, IRI__GEOSPARQL__HAS_GEOMETRY, locations);
+      _writer->writeTriple(subj, IRI__GEOSPARQL__HAS_GEOMETRY, geomObj);
+      writeBoostGeometry(geomObj, IRI__GEOSPARQL__AS_WKT, locations);
+    } else {
+      writeBoostGeometry(subj, IRI__GEOSPARQL__HAS_GEOMETRY, locations);
+    }
   }
 
   if (_config.addWayConvexHull) {