feat: Handle missing ground truth in dataset

cpp/bindings/types.h

@@ -267,6 +267,13 @@ inline void make_types(nb::module_& m) {
       "List of points in the "
       "point cloud. Note, this is always in row major format."
     )
+    .def_static(
+      "from_vec_positions",
+      &LidarMeasurement::from_vec_positions,
+      "stamp"_a,
+      "positions"_a,
+      "Construct a LidarMeasurement from a stamp and a (n,3) numpy array."
+    )
     .def(
       "to_vec_positions",
       &LidarMeasurement::to_vec_positions,

cpp/evalio/types.h

@@ -218,6 +218,16 @@ struct LidarMeasurement {
     return oss.str();
   }
 
+  static LidarMeasurement
+  from_vec_positions(Stamp stamp, const Eigen::MatrixX3d& positions) {
+    std::vector<Point> pts;
+    pts.reserve(positions.rows());
+    for (const auto& p : positions.rowwise()) {
+      pts.push_back(Point {.x = p.x(), .y = p.y(), .z = p.z()});
+    }
+    return LidarMeasurement(stamp, pts);
+  }
+
   std::vector<Eigen::Vector3d> to_vec_positions() const {
     std::vector<Eigen::Vector3d> eigen_points;
     eigen_points.reserve(points.size());

pyproject.toml

@@ -41,6 +41,14 @@ build-backend = "scikit_build_core.build"
 evalio = "evalio.cli:app"
 
 # -------------- Tools -------------- #
+[tool.uv]
+cache-keys = [
+    { file = "pyproject.toml" },
+    { file = "cpp/**/*.{h,c,hpp,cpp}" },
+    { file = "python/**/*.{cpp,py}" },
+    { file = "CMakeLists.txt" },
+]
+
 # building
 [tool.scikit-build]
 minimum-version = "0.12"
@@ -49,6 +57,7 @@ sdist.include = ["cpp/bindings/pipelines-src/*"]
 
 [tool.scikit-build.cmake.define]
 EVALIO_BUILD_PYTHON = true
+CMAKE_BUILD_TYPE = "RelWithDebInfo"
 VCPKG_INSTALLED_DIR = "./.vcpkg_installed"
 # VCPKG_INSTALL_OPTIONS = "--debug"
 

python/evalio/cli/ls.py

@@ -142,7 +142,9 @@ def ls(
                 )
             else:
                 # sequences
-                all_info["Sequences"].append("\n".join(d.sequences()))
+                all_info["Sequences"].append(
+                    "\n".join([v.seq_name for v in d.sequences()])
+                )
                 # downloaded
                 downloaded = [d(s).is_downloaded() for s in d.sequences()]
                 downloaded = "\n".join(

python/evalio/cli/run.py

@@ -274,7 +274,9 @@ def run(
 
         # save ground truth if we haven't already
         if not (gt_file := exp.file.parent / "gt.csv").exists():
-            exp.sequence.ground_truth().to_file(gt_file)
+            gt = exp.sequence.ground_truth()
+            if gt is not None:
+                gt.to_file(gt_file)
 
         # Figure out the status of the experiment
         traj = ty.Trajectory.from_file(exp.file)

python/evalio/cli/stats.py

@@ -43,10 +43,6 @@ def eval_dataset(
         elif isinstance(traj.metadata, ty.Experiment):
             all_trajs.append(cast(ty.Trajectory[ty.Experiment], traj))
 
-    if gt_og is None:
-        print_warning(f"No ground truth found in {dir}, skipping.")
-        return None
-
     # Setup visualization
     if visualize:
         try:
@@ -67,11 +63,12 @@ def eval_dataset(
             spawn=False,
         )
         rr.connect_grpc()
-        rr.log(
-            "gt",
-            convert(gt_og, color=GT_COLOR),
-            static=True,
-        )
+        if gt_og is not None:
+            rr.log(
+                "gt",
+                convert(gt_og, color=GT_COLOR),
+                static=True,
+            )
 
         # generate colors for visualization
         colors = distinctipy.get_colors(len(all_trajs) + 1)
@@ -86,7 +83,7 @@ def eval_dataset(
         if traj.metadata.total_elapsed is not None:
             hz = len(traj) / traj.metadata.total_elapsed
 
-        if len(traj) > 0:
+        if len(traj) > 0 and gt_og is not None:
             # align to ground truth, copying ground truth by hand
             gt_aligned = Trajectory(
                 stamps=[ty.Stamp(s) for s in gt_og.stamps],
@@ -296,7 +293,7 @@ def evaluate_typer(
     else:
         from asteval import Interpreter
 
-        filter_method = lambda r: Interpreter(user_symbols=r).eval(
+        filter_method = lambda r: Interpreter(user_symbols=r).eval(  # type: ignore
             filter_str, raise_errors=True
         )
 

python/evalio/datasets/base.py

@@ -84,14 +84,14 @@ def data_iter(self) -> DatasetIterator:
         ...
 
     # Return the ground truth in the ground truth frame
-    def ground_truth_raw(self) -> Trajectory:
+    def ground_truth_raw(self) -> Optional[Trajectory]:
         """
         Retrieves the raw ground truth trajectory, as represented in the ground truth frame.
 
         Returns:
             The raw ground truth trajectory data.
         """
-        ...
+        return None
 
     # ------------------------- For loading params ------------------------- #
     def imu_T_lidar(self) -> SE3:
@@ -210,13 +210,16 @@ def is_downloaded(self) -> bool:
 
         return True
 
-    def ground_truth(self) -> Trajectory[GroundTruth]:
+    def ground_truth(self) -> Optional[Trajectory[GroundTruth]]:
         """Get the ground truth trajectory in the **IMU** frame, rather than the ground truth frame as returned in [ground_truth_raw][evalio.datasets.Dataset.ground_truth_raw].
 
         Returns:
             The ground truth trajectory in the IMU frame.
         """
         gt_traj = self.ground_truth_raw()
+        if gt_traj is None:
+            return None
+
         gt_T_imu = self.imu_T_gt().inverse()
 
         # Convert to IMU frame

python/evalio/datasets/loaders.py

@@ -368,8 +368,8 @@ def _step(iter: Iterator[T]) -> Optional[T]:
             return None
 
     def __iter__(self) -> Iterator[Measurement]:
-        self.next_imu = next(self.iter_imu)
-        self.next_lidar = next(self.iter_lidar)
+        self.next_imu = self._step(self.iter_imu)
+        self.next_lidar = self._step(self.iter_lidar)
         return self
 
     def __next__(self) -> Measurement:

python/evalio/rerun.py

@@ -127,7 +127,8 @@ def new_dataset(self, dataset: Dataset):
             self.lidar_params = dataset.lidar_params()
             self.imu_T_lidar = dataset.imu_T_lidar()
 
-            self.rec.log("gt", convert(self.gt, color=GT_COLOR), static=True)
+            if self.gt is not None:
+                self.rec.log("gt", convert(self.gt, color=GT_COLOR), static=True)
 
             # reset other variables
             self.pn = None
@@ -165,7 +166,7 @@ def log_scan(self, data: LidarMeasurement):
             if self.args is None:
                 return
 
-            if self.lidar_params is None or self.gt is None:
+            if self.lidar_params is None:
                 raise ValueError("You needed to add a dataset before stepping!")
             if self.pn is None or self.trajectory is None or self.colors is None:
                 raise ValueError("You needed to add a pipeline before stepping!")
@@ -192,13 +193,13 @@ def log_pose(self, stamp: Stamp, pose: SE3):
             if self.args is None:
                 return
 
-            if self.lidar_params is None or self.gt is None:
+            if self.lidar_params is None:
                 raise ValueError("You needed to add a dataset before stepping!")
             if self.pn is None or self.trajectory is None or self.colors is None:
                 raise ValueError("You needed to add a pipeline before stepping!")
 
             # Find transform between ground truth and imu origins
-            if self.gt_o_T_imu_o is None:
+            if self.gt_o_T_imu_o is None and self.gt is not None:
                 # If we haven't hit ground truth data or the pose isn't being updated, wait
                 if stamp < self.gt.stamps[0] or pose == SE3.identity():
                     pass
@@ -231,7 +232,7 @@ def log_map(self, stamp: Stamp, map: dict[str, NDArray[np.float64]]):
             if self.args is None:
                 return
 
-            if self.lidar_params is None or self.gt is None:
+            if self.lidar_params is None:
                 raise ValueError("You needed to add a pipeline before stepping!")
             if self.pn is None or self.trajectory is None or self.colors is None:
                 raise ValueError("You needed to add a pipeline before stepping!")
@@ -247,7 +248,7 @@ def log_features(self, stamp: Stamp, features: dict[str, NDArray[np.float64]]):
             if self.args is None:
                 return
 
-            if self.lidar_params is None or self.gt is None:
+            if self.lidar_params is None:
                 raise ValueError("You needed to add a pipeline before stepping!")
             if self.pn is None or self.trajectory is None or self.colors is None:
                 raise ValueError("You needed to add a pipeline before stepping!")