feat: add scalar WKB reader/writer

CMakeLists.txt

@@ -207,6 +207,7 @@ add_library(s2geography
     src/s2geography/geoarrow.cc
     src/s2geography/wkt-reader.cc
     src/s2geography/wkt-writer.cc
+    src/s2geography/wkb.cc
     src/s2geography/op/cell.cc
     src/s2geography/op/point.cc
     # geoarrow sources
@@ -259,6 +260,7 @@ if(S2GEOGRAPHY_BUILD_TESTS)
   add_executable(geoarrow_test src/s2geography/geoarrow_test.cc)
   add_executable(geography_test src/s2geography/geography_test.cc)
   add_executable(wkt_writer_test src/s2geography/wkt-writer_test.cc)
+  add_executable(wkb_test src/s2geography/wkb_test.cc)
   add_executable(op_cell_test src/s2geography/op/cell_test.cc)
 
   if (S2GEOGRAPHY_CODE_COVERAGE)
@@ -271,6 +273,7 @@ if(S2GEOGRAPHY_BUILD_TESTS)
   target_link_libraries(geoarrow_test s2geography nanoarrow GTest::gtest_main GTest::gmock)
   target_link_libraries(geography_test s2geography nanoarrow GTest::gtest_main)
   target_link_libraries(wkt_writer_test s2geography GTest::gtest_main)
+  target_link_libraries(wkb_test s2geography GTest::gtest_main)
   target_link_libraries(op_cell_test s2geography GTest::gtest_main)
 
   target_include_directories(geoarrow_test PRIVATE src/vendored)
@@ -280,6 +283,7 @@ if(S2GEOGRAPHY_BUILD_TESTS)
   gtest_discover_tests(geoarrow_test)
   gtest_discover_tests(geography_test)
   gtest_discover_tests(wkt_writer_test)
+  gtest_discover_tests(wkb_test)
   gtest_discover_tests(op_cell_test)
 endif()
 

src/s2geography.h

@@ -18,3 +18,4 @@
 #include "s2geography/projections.h"
 #include "s2geography/wkt-reader.h"
 #include "s2geography/wkt-writer.h"
+#include "s2geography/wkb.h"

src/s2geography/wkb.cc

@@ -0,0 +1,67 @@
+
+#include "s2geography/wkb.h"
+
+#include <cstdlib>
+#include <cstring>
+#include <sstream>
+
+#include "s2geography/geoarrow.h"
+#include "s2geography/geography.h"
+
+namespace s2geography {
+
+WKBReader::WKBReader(const geoarrow::ImportOptions& options) {
+  reader_ = absl::make_unique<geoarrow::Reader>();
+  reader_->Init(geoarrow::Reader::InputType::kWKB, options);
+}
+
+std::unique_ptr<Geography> WKBReader::ReadFeature(const uint8_t* bytes,
+                                                  int64_t size) {
+  if (size > std::numeric_limits<int32_t>::max()) {
+    throw Exception("Can't parse WKB greater than 2GB in size");
+  }
+
+  int32_t offsets[] = {0, static_cast<int32_t>(size)};
+  const void* buffers[] = {nullptr, offsets, bytes};
+  ArrowArray array;
+  array.length = 1;
+  array.null_count = 0;
+  array.offset = 0;
+
+  array.n_buffers = 3;
+  array.n_children = 0;
+  array.buffers = buffers;
+
+  array.children = nullptr;
+  array.dictionary = nullptr;
+  array.release = [](ArrowArray*) -> void {};
+  array.private_data = nullptr;
+
+  out_.clear();
+  reader_->ReadGeography(&array, 0, 1, &out_);
+  return std::move(out_[0]);
+}
+
+std::unique_ptr<Geography> WKBReader::ReadFeature(const std::string_view bytes) {
+  return ReadFeature(reinterpret_cast<const uint8_t*>(bytes.data()), bytes.size());
+}
+
+WKBWriter::WKBWriter(const geoarrow::ExportOptions& options) {
+  writer_ = absl::make_unique<geoarrow::Writer>();
+  writer_->Init(geoarrow::Writer::OutputType::kWKB, options);
+
+}
+
+std::string WKBWriter::WriteFeature(const Geography& geog) {
+  ArrowArray array;
+  writer_->WriteGeography(geog);
+  writer_->Finish(&array);
+
+  const auto offsets = static_cast<const int32_t*>(array.buffers[1]);
+  const auto data = static_cast<const uint8_t*>(array.buffers[2]);
+  std::string result(reinterpret_cast<const char*>(data), offsets[1]);
+
+  return result;
+}
+
+}  // namespace s2geography

src/s2geography/wkb.h

@@ -0,0 +1,33 @@
+
+#pragma once
+
+#include "s2geography/geoarrow.h"
+#include "s2geography/geography.h"
+
+namespace s2geography {
+
+class WKBReader {
+ public:
+  WKBReader() : WKBReader(geoarrow::ImportOptions()) {}
+  WKBReader(const geoarrow::ImportOptions& options);
+
+  std::unique_ptr<Geography> ReadFeature(const uint8_t* bytes, int64_t size);
+  std::unique_ptr<Geography> ReadFeature(const std::string_view bytes);
+
+ private:
+  std::unique_ptr<geoarrow::Reader> reader_;
+  std::vector<std::unique_ptr<Geography>> out_;
+};
+
+class WKBWriter {
+ public:
+  WKBWriter() : WKBWriter(geoarrow::ExportOptions()) {}
+  WKBWriter(const geoarrow::ExportOptions& options);
+
+  std::string WriteFeature(const Geography& geog);
+
+ private:
+  std::unique_ptr<geoarrow::Writer> writer_;
+};
+
+}  // namespace s2geography

src/s2geography/wkb_test.cc

@@ -0,0 +1,75 @@
+
+#include <gtest/gtest.h>
+
+#include "s2geography.h"
+
+using namespace s2geography;
+
+void wkbRoundTrip(const std::string wktIn,
+                  const std::vector<uint8_t> &wkbIn) {
+  WKTWriter wkt_writer(2);
+  WKBReader reader;
+  WKBWriter writer;
+
+  auto geog = reader.ReadFeature(wkbIn.data(), wkbIn.size());
+  auto wkbOut = writer.WriteFeature(*geog);
+  auto geogOut = reader.ReadFeature(wkbOut);
+  std::string wktOut = wkt_writer.write_feature(*geogOut);
+  EXPECT_EQ(wktOut, wktIn);
+}
+
+TEST(WKBRoundtrip, Point) {
+  std::string wkt("POINT (30 10)");
+  std::vector<uint8_t> wkb({0x01, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00,
+                            0x00, 0x00, 0x00, 0x00, 0x3e, 0x40, 0x00,
+                            0x00, 0x00, 0x00, 0x00, 0x00, 0x24, 0x40});
+  wkbRoundTrip(wkt, wkb);
+}
+
+TEST(WKBRoundtrip, LineString) {
+  std::string wkt("LINESTRING (30 10, 12 42)");
+  std::vector<uint8_t> wkb(
+      {0x01, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00,
+       0x00, 0x00, 0x00, 0x00, 0x3e, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00,
+       0x00, 0x24, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x28, 0x40,
+       0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x45, 0x40});
+  wkbRoundTrip(wkt, wkb);
+}
+
+TEST(WKBRoundtrip, Polygon) {
+  std::string wkt(
+      "POLYGON ((35 10, 45 45, 15 40, 10 20, 35 10), (20 30, 35 35, 30 20, 20 "
+      "30))");
+  std::vector<uint8_t> wkb(
+      {0x01, 0x03, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00,
+       0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, 0x41, 0x40, 0x00, 0x00, 0x00,
+       0x00, 0x00, 0x00, 0x24, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, 0x46,
+       0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, 0x46, 0x40, 0x00, 0x00, 0x00,
+       0x00, 0x00, 0x00, 0x2e, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x44,
+       0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x24, 0x40, 0x00, 0x00, 0x00,
+       0x00, 0x00, 0x00, 0x34, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, 0x41,
+       0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x24, 0x40, 0x04, 0x00, 0x00,
+       0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x34, 0x40, 0x00, 0x00, 0x00,
+       0x00, 0x00, 0x00, 0x3e, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, 0x41,
+       0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, 0x41, 0x40, 0x00, 0x00, 0x00,
+       0x00, 0x00, 0x00, 0x3e, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x34,
+       0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x34, 0x40, 0x00, 0x00, 0x00,
+       0x00, 0x00, 0x00, 0x3e, 0x40});
+  wkbRoundTrip(wkt, wkb);
+}
+
+TEST(WKBRoundtrip, ExportOptions) {
+  s2geography::WKTReader reader;
+  auto geog = reader.read_feature("LINESTRING (-64 45, 0 45)");
+
+  WKBWriter writer;
+  auto wkbOut = writer.WriteFeature(*geog);
+
+  // smoke test for passing through options: with tessellation -> more coordinates
+  s2geography::geoarrow::ExportOptions options;
+  options.set_tessellate_tolerance(S1Angle::Radians(0.001));
+  WKBWriter writer2(options);
+  auto wkbOut2 = writer2.WriteFeature(*geog);
+
+  EXPECT_GT(wkbOut2.size(), wkbOut.size());
+}