pro_rrt::scenario_runner Namespace Reference

Classes
struct	RunResult

Functions
tl::expected< RunResult, std::string >	runScenario (const std::filesystem::path &scenario_dir)

Detailed Description

In-process scenario runner. Reads a captured pro_rrt scenario from disk, reconstructs the planner inputs (robot model, planning scene, request), and re-runs pro_rrt::planTrajectoryToJointGoal — without spawning a subprocess. Both the standalone pro_rrt_replay binary and the in-process pro_rrt_benchmark driver call this.

The caller must have already invoked rclcpp::init() exactly once at process startup, and is responsible for rclcpp::shutdown() at exit. Each call to runScenario creates a transient rclcpp::Node (with the scenario's joint_limits.yaml / kinematics.yaml loaded via NodeOptions::arguments({"--ros-args","--params-file",…})) and destroys it on return; no per-scenario state leaks between calls.

Function Documentation

runScenario()

tl::expected< RunResult, std::string > pro_rrt::scenario_runner::runScenario

(

const std::filesystem::path &

scenario_dir

)

Returns tl::unexpected with a human-readable message on disk / parse / setup failure. The planner returning a failure (e.g. "no path found") is reported as result.success=false inside the value path, NOT as an unexpected.