feat(dc_measurements): add the Mission Measurement Open-RMF adapter

[Minipada]

Summary

Adds mission_open_rmf (class MissionOpenRmf), the Open-RMF adapter of the Mission
Measurement: it consumes Open-RMF's TaskState stream over a plain ws:// connection
(dc_common::WebSocketJsonClient, #390) and emits mission_start/mission_end Records in
#305/#387's contract — mission_id from booking.id, mission_type from category.

Open-RMF's task model is a continuously-updated 12-value status enum rather than a flat
start/end pair; the status names and TaskState field shapes used here (booking, category,
cancellation.labels, killed.labels, dispatch.errors, unix_millis_start_time/
unix_millis_finish_time) were verified against the real rmf_api_msgs task_state.json
schema, not guessed.

Per the acceptance criteria, two judgment calls are resolved explicitly (see
MissionOpenRmfCore's doc comment and doc/src/dc/measurements/mission_open_rmf.md) rather than
silently defaulted:

• skipped is treated as task-terminal, mapping to outcome cancelled (DC's outcome contract
  has four values, not five).
• blocked/error are treated as transient, not terminal — the task stays open until a real
  terminal status arrives.

reason/error_code are populated best-effort from whichever field Open-RMF actually supplied
(unlike nav2's always-present error_msg), and duration_sec prefers Open-RMF's own
unix_millis_start_time/unix_millis_finish_time over locally observed timestamps when both are
present.

Note on #305: this issue lists #305 (the still-open mission-lifecycle-contract design issue)
as a blocker. #388/#389, which share that same dependency, already implemented and merged against
the same mission_start/mission_end contract, so this PR follows that established precedent
rather than waiting on #305's tracking issue to be formally closed.

Note on build verification: local disk space constraints prevented a full colcon build/colcon test run of the whole workspace in this environment. The new ROS-free
MissionOpenRmfCore logic (the actual status-mapping/outcome decisions) was independently
compiled and unit-tested (16/16 passing) with g++ against gtest directly. dc_common,
dc_lifecycle_manager, and dc_bridge all compiled successfully during the interrupted
workspace build attempts, so CI's clean-runner colcon build/colcon test is the outstanding
signal for the full plugin + integration test.

Closes #391

Test plan

• MissionOpenRmfCore unit tests (test_mission_open_rmf_core.cpp, 16 cases) — compiled and
  run standalone with g++/gtest, all passing
• prek run --all-files --skip build-doc — clean
• CI: colcon build + colcon test for dc_measurements (new plugin + integration test
  test_measurement_mission_open_rmf.cpp against a mock websocket TaskState fixture)

🤖 Generated with Claude Code

https://claude.ai/code/session_01Dsswe9mDB56HKL5x9igcma

[codecov]

Codecov Report

❌ Patch coverage is 81.73516% with 40 lines in your changes missing coverage. Please review.
✅ Project coverage is 70.80%. Comparing base (8ba8a53) to head (e2ee604).

Files with missing lines	Patch %	Lines
...urements/plugins/measurements/mission_open_rmf.cpp	84.06%	22 Missing ⚠️
...nts/plugins/measurements/mission_open_rmf_core.hpp	77.78%	18 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##            jazzy     #417      +/-   ##
==========================================
+ Coverage   70.47%   70.80%   +0.34%     
==========================================
  Files         112      114       +2     
  Lines        7221     7440     +219     
==========================================
+ Hits         5088     5267     +179     
- Misses       2133     2173      +40     

Flag	Coverage Δ
cpp-jazzy	70.80% <81.74%> (+0.34%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.