feat: add autoware_lanelet2_map_merger package by mitsudome-r · Pull Request #410 · autowarefoundation/autoware_tools

mitsudome-r · 2026-04-22T07:52:23Z

Description

Part of autowarefoundation/autoware_core#887

Add a new autoware_lanelet2_map_merger package that merges a directory of divided Lanelet2 (.osm) map files into a single .osm file.

New ROS 2 node autoware_lanelet2_map_merger_node with launch file, parameter YAML, and JSON schema.
Loads every .osm file in the input directory using the projector described in map_projector_info.yaml (MGRS, LocalCartesianUTM, LocalCartesian, and TransverseMercator are supported).
Merges the loaded maps using the same logic as lanelet2_map_loader (merge_lanelet2_maps), deduplicating points shared across maps by ID so successor/predecessor relationships are preserved, then writes the merged map with lanelet::write.
Includes README and a gtest covering the merge behavior.

How was this PR tested?

Built the package with colcon build --packages-select autoware_lanelet2_map_merger.
Ran the included gtest suite (test_lanelet2_map_merger) via colcon test.
Verified end-to-end round-trip by running autoware_lanelet2_map_divider on a sample lanelet2_map.osm and then running this merger on its output, confirming the reconstructed map matches the original.

Notes for reviewers

None.

Effects on system behavior

None.

Signed-off-by: Ryohsuke Mitsudome <ryohsuke.mitsudome@tier4.jp>

github-actions · 2026-04-22T07:52:34Z

Thank you for contributing to the Autoware project!

🚧 If your pull request is in progress, switch it to draft mode.

Please ensure:

You've checked our contribution guidelines.
Your PR follows our pull request guidelines.
All required CI checks pass before marking the PR ready for review.

Signed-off-by: Ryohsuke Mitsudome <ryohsuke.mitsudome@tier4.jp>

Copilot

Pull request overview

Adds a new ROS 2 “map tool” package that merges a directory of divided Lanelet2 .osm files back into a single .osm, using a projector described by map_projector_info.yaml and preserving shared-point IDs to keep topology consistent.

Changes:

Introduces autoware_lanelet2_map_merger library + component node to discover, load, merge, and write Lanelet2 maps.
Adds launch/config + parameter JSON schema and README documentation.
Adds a gtest that exercises merge/dedup and reload behavior.

Reviewed changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
map/autoware_lanelet2_map_merger/src/lanelet2_map_merger.cpp	Core discovery/load/merge/write implementation.
map/autoware_lanelet2_map_merger/src/lanelet2_map_merger.hpp	Merger class interface and state (loaded map lifetime).
map/autoware_lanelet2_map_merger/src/lanelet2_map_merger_node.cpp	One-shot ROS 2 node wrapper that runs the merger from parameters.
map/autoware_lanelet2_map_merger/src/lanelet2_map_merger_node.hpp	Node class declaration.
map/autoware_lanelet2_map_merger/src/local_projector.hpp	Minimal projector used for `LOCAL` projector-type handling.
map/autoware_lanelet2_map_merger/test/test_lanelet2_map_merger.cpp	Unit tests for merge/dedup + reload + non-crash behavior.
map/autoware_lanelet2_map_merger/CMakeLists.txt	Build/test targets and component registration.
map/autoware_lanelet2_map_merger/package.xml	Package metadata and dependencies.
map/autoware_lanelet2_map_merger/launch/lanelet2_map_merger.launch.xml	Launch entrypoint wiring parameters.
map/autoware_lanelet2_map_merger/config/lanelet2_map_merger.param.yaml	Default parameter file using launch substitutions.
map/autoware_lanelet2_map_merger/schema/lanelet2_map_merger.schema.json	Parameter schema for documentation/validation.
map/autoware_lanelet2_map_merger/README.md	Usage and behavior documentation.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-04-22T08:48:00Z

+void Lanelet2MapMerger::run()
+{
+  const auto osm_files = discover_osm_files(input_dir_);
+  if (osm_files.empty()) {
+    RCLCPP_ERROR(logger_, "No .osm files found under %s", input_dir_.c_str());
+    return;
+  }


loaded_maps_ is an instance member used to keep source maps alive, but it is never cleared. If run() is called more than once on the same Lanelet2MapMerger instance, maps from previous runs will accumulate, increasing memory usage unnecessarily. Clearing loaded_maps_ at the start of run() (or before loading a new set of maps) avoids this.

Copilot · 2026-04-22T08:48:01Z

+  for (auto ls : {ls_left_a, ls_right_a, ls_right_b}) {
+    ls.attributes()["type"] = "line_thin";
+    ls.attributes()["subtype"] = "solid";
+  }
+
+  lanelet::Lanelet ll_a(100, ls_left_a, ls_right_a);
+  lanelet::Lanelet ll_b(101, ls_right_a, ls_right_b);
+  for (auto ll : {ll_a, ll_b}) {
+    ll.attributes()["type"] = "lanelet";
+    ll.attributes()["subtype"] = "road";
+    ll.attributes()["location"] = "urban";
+    ll.attributes()["one_way"] = "yes";
+  }


In make_sample_map(), this loop iterates the line strings by value (for (auto ls : {ls_left_a, ...})), so the attribute assignments are applied to temporary copies and do not affect ls_left_a/ls_right_a/ls_right_b. If these attributes are intended to be written into the generated .osm, iterate by reference (e.g., for (auto & ls : ...)) or set attributes directly on each linestring.

Suggested change

for (auto ls : {ls_left_a, ls_right_a, ls_right_b}) {

ls.attributes()["type"] = "line_thin";

ls.attributes()["subtype"] = "solid";

}

lanelet::Lanelet ll_a(100, ls_left_a, ls_right_a);

lanelet::Lanelet ll_b(101, ls_right_a, ls_right_b);

for (auto ll : {ll_a, ll_b}) {

ll.attributes()["type"] = "lanelet";

ll.attributes()["subtype"] = "road";

ll.attributes()["location"] = "urban";

ll.attributes()["one_way"] = "yes";

}

ls_left_a.attributes()["type"] = "line_thin";

ls_left_a.attributes()["subtype"] = "solid";

ls_right_a.attributes()["type"] = "line_thin";

ls_right_a.attributes()["subtype"] = "solid";

ls_right_b.attributes()["type"] = "line_thin";

ls_right_b.attributes()["subtype"] = "solid";

lanelet::Lanelet ll_a(100, ls_left_a, ls_right_a);

lanelet::Lanelet ll_b(101, ls_right_a, ls_right_b);

ll_a.attributes()["type"] = "lanelet";

ll_a.attributes()["subtype"] = "road";

ll_a.attributes()["location"] = "urban";

ll_a.attributes()["one_way"] = "yes";

ll_b.attributes()["type"] = "lanelet";

ll_b.attributes()["subtype"] = "road";

ll_b.attributes()["location"] = "urban";

ll_b.attributes()["one_way"] = "yes";

Copilot · 2026-04-22T08:48:01Z

+  for (auto ls : {ls_left_a, ls_right_a, ls_right_b}) {
+    ls.attributes()["type"] = "line_thin";
+    ls.attributes()["subtype"] = "solid";
+  }
+
+  lanelet::Lanelet ll_a(100, ls_left_a, ls_right_a);
+  lanelet::Lanelet ll_b(101, ls_right_a, ls_right_b);
+  for (auto ll : {ll_a, ll_b}) {
+    ll.attributes()["type"] = "lanelet";
+    ll.attributes()["subtype"] = "road";
+    ll.attributes()["location"] = "urban";
+    ll.attributes()["one_way"] = "yes";
+  }


Similarly, this loop iterates lanelets by value (for (auto ll : {ll_a, ll_b})), so the lanelet attribute assignments won't be applied to ll_a/ll_b before adding them to the map. If the test intends to produce realistic lanelet relations (type/subtype/etc.), iterate by reference or set the attributes on ll_a and ll_b directly.

Suggested change

for (auto ls : {ls_left_a, ls_right_a, ls_right_b}) {

ls.attributes()["type"] = "line_thin";

ls.attributes()["subtype"] = "solid";

}

lanelet::Lanelet ll_a(100, ls_left_a, ls_right_a);

lanelet::Lanelet ll_b(101, ls_right_a, ls_right_b);

for (auto ll : {ll_a, ll_b}) {

ll.attributes()["type"] = "lanelet";

ll.attributes()["subtype"] = "road";

ll.attributes()["location"] = "urban";

ll.attributes()["one_way"] = "yes";

}

ls_left_a.attributes()["type"] = "line_thin";

ls_left_a.attributes()["subtype"] = "solid";

ls_right_a.attributes()["type"] = "line_thin";

ls_right_a.attributes()["subtype"] = "solid";

ls_right_b.attributes()["type"] = "line_thin";

ls_right_b.attributes()["subtype"] = "solid";

lanelet::Lanelet ll_a(100, ls_left_a, ls_right_a);

lanelet::Lanelet ll_b(101, ls_right_a, ls_right_b);

ll_a.attributes()["type"] = "lanelet";

ll_a.attributes()["subtype"] = "road";

ll_a.attributes()["location"] = "urban";

ll_a.attributes()["one_way"] = "yes";

ll_b.attributes()["type"] = "lanelet";

ll_b.attributes()["subtype"] = "road";

ll_b.attributes()["location"] = "urban";

ll_b.attributes()["one_way"] = "yes";

Copilot · 2026-04-22T08:48:01Z

+    tmp_dir_ = fs::temp_directory_path() / "test_lanelet2_map_merger";
+    fs::remove_all(tmp_dir_);
+    fs::create_directories(tmp_dir_);
+  }


The test uses a fixed temp directory name (fs::temp_directory_path() / "test_lanelet2_map_merger"). If tests are run in parallel (e.g., ctest -j), multiple instances can race on remove_all/create_directories, causing flaky failures. Prefer creating a unique per-test directory (PID + random suffix) and cleaning it up in TearDown().

Copilot · 2026-04-22T08:48:02Z

+if(BUILD_TESTING)
+  find_package(ament_lint_auto REQUIRED)
+  ament_lint_auto_find_test_dependencies()
+
+  ament_add_gtest(test_lanelet2_map_merger test/test_lanelet2_map_merger.cpp)
+  target_link_libraries(test_lanelet2_map_merger ${PROJECT_NAME})
+endif()


The test block calls ament_add_gtest(...) but does not find_package(ament_cmake_gtest REQUIRED) first. In this repo, other packages that use ament_add_gtest explicitly find ament_cmake_gtest in the BUILD_TESTING block; without it, CMake configuration can fail because the macro isn't defined.

Copilot · 2026-04-22T08:48:02Z

+    if (is_local) {
+      for (lanelet::Point3d point : map->pointLayer) {
+        if (point.hasAttribute("local_x")) {
+          point.x() = point.attribute("local_x").asDouble().value();
+        }
+        if (point.hasAttribute("local_y")) {
+          point.y() = point.attribute("local_y").asDouble().value();
+        }
+      }


In the LOCAL-projector path, the loop iterates map->pointLayer by value (for (lanelet::Point3d point : ...)). Any updates to point.x()/point.y() are applied to the temporary copy and do not modify the points stored in the map, so the merged output will keep the (projector-generated) coordinates (likely all zeros). Iterate by reference (e.g., auto & point) so the loaded map geometry is actually updated before merging/writing.

Copilot · 2026-04-22T08:48:02Z

+          point.x() = point.attribute("local_x").asDouble().value();
+        }
+        if (point.hasAttribute("local_y")) {
+          point.y() = point.attribute("local_y").asDouble().value();


point.attribute("local_x").asDouble().value() (and the same for local_y) will throw if the attribute exists but cannot be parsed as a double. Consider checking the optional returned by asDouble() and logging/handling invalid values to avoid an unexpected exception when merging maps with malformed tags.

Suggested change

point.x() = point.attribute("local_x").asDouble().value();

}

if (point.hasAttribute("local_y")) {

point.y() = point.attribute("local_y").asDouble().value();

const auto local_x = point.attribute("local_x").asDouble();

if (!local_x) {

RCLCPP_ERROR(

logger_, "Invalid local_x attribute in %s for point id %ld", path.c_str(), point.id());

return nullptr;

}

point.x() = *local_x;

}

if (point.hasAttribute("local_y")) {

const auto local_y = point.attribute("local_y").asDouble();

if (!local_y) {

RCLCPP_ERROR(

logger_, "Invalid local_y attribute in %s for point id %ld", path.c_str(), point.id());

return nullptr;

}

point.y() = *local_y;

feat: add autoware_lanelet2_map_merger package

4ec7ff0

Signed-off-by: Ryohsuke Mitsudome <ryohsuke.mitsudome@tier4.jp>

github-actions Bot added type:documentation component:map labels Apr 22, 2026

style(pre-commit): autofix

008a2d1

mitsudome-r added the tag:run-build-and-test-differential label Apr 22, 2026

mitsudome-r mentioned this pull request Apr 22, 2026

Implement Differential Map Loading feature for lanelet2 map autowarefoundation/autoware_core#887

Open

11 tasks

mitsudome-r requested review from YamatoAndo, Copilot and maxime-clem April 22, 2026 08:41

Copilot started reviewing on behalf of mitsudome-r April 22, 2026 08:41 View session

fix spell check error

2fd7cd3

Signed-off-by: Ryohsuke Mitsudome <ryohsuke.mitsudome@tier4.jp>

Copilot AI reviewed Apr 22, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: add autoware_lanelet2_map_merger package#410

feat: add autoware_lanelet2_map_merger package#410
mitsudome-r wants to merge 3 commits intoautowarefoundation:mainfrom
mitsudome-r:feat/add-autoware-lanelet2-map-merger

mitsudome-r commented Apr 22, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Apr 22, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

-  for (auto ls : {ls_left_a, ls_right_a, ls_right_b}) {
-    ls.attributes()["type"] = "line_thin";
-    ls.attributes()["subtype"] = "solid";
-  }
-  lanelet::Lanelet ll_a(100, ls_left_a, ls_right_a);
-  lanelet::Lanelet ll_b(101, ls_right_a, ls_right_b);
-  for (auto ll : {ll_a, ll_b}) {
-    ll.attributes()["type"] = "lanelet";
-    ll.attributes()["subtype"] = "road";
-    ll.attributes()["location"] = "urban";
-    ll.attributes()["one_way"] = "yes";
-  }
+  ls_left_a.attributes()["type"] = "line_thin";
+  ls_left_a.attributes()["subtype"] = "solid";
+  ls_right_a.attributes()["type"] = "line_thin";
+  ls_right_a.attributes()["subtype"] = "solid";
+  ls_right_b.attributes()["type"] = "line_thin";
+  ls_right_b.attributes()["subtype"] = "solid";
+  lanelet::Lanelet ll_a(100, ls_left_a, ls_right_a);
+  lanelet::Lanelet ll_b(101, ls_right_a, ls_right_b);
+  ll_a.attributes()["type"] = "lanelet";
+  ll_a.attributes()["subtype"] = "road";
+  ll_a.attributes()["location"] = "urban";
+  ll_a.attributes()["one_way"] = "yes";
+  ll_b.attributes()["type"] = "lanelet";
+  ll_b.attributes()["subtype"] = "road";
+  ll_b.attributes()["location"] = "urban";
+  ll_b.attributes()["one_way"] = "yes";

-          point.x() = point.attribute("local_x").asDouble().value();
-        }
-        if (point.hasAttribute("local_y")) {
-          point.y() = point.attribute("local_y").asDouble().value();
+          const auto local_x = point.attribute("local_x").asDouble();
+          if (!local_x) {
+            RCLCPP_ERROR(
+              logger_, "Invalid local_x attribute in %s for point id %ld", path.c_str(), point.id());
+            return nullptr;
+          }
+          point.x() = *local_x;
+        }
+        if (point.hasAttribute("local_y")) {
+          const auto local_y = point.attribute("local_y").asDouble();
+          if (!local_y) {
+            RCLCPP_ERROR(
+              logger_, "Invalid local_y attribute in %s for point id %ld", path.c_str(), point.id());
+            return nullptr;
+          }
+          point.y() = *local_y;

Conversation

mitsudome-r commented Apr 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

How was this PR tested?

Notes for reviewers

Effects on system behavior

Uh oh!

github-actions Bot commented Apr 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

mitsudome-r commented Apr 22, 2026 •

edited

Loading

github-actions Bot commented Apr 22, 2026 •

edited

Loading