Add a rotor_scan_resolution input key

[alongd]

Motivation

ARC had no user-facing way to state the resolution of its 1D rotor scans. The value came from
rotor_scan_resolution in the settings of whatever process happened to be running. This bit us
for real: a machine-wide settings override silently put every rotor scan on a workstation below
the resolution where the downstream Fourier fit is valid, for a week, with no crash and no
warning — a silent wrong answer. This PR makes a run able to state what it needs, so resolution is
a property of the run and not of the host that launched it.

What changed

Adds a user-facing input key rotor_scan_resolution (degrees; points per rotor = 360 / value):

project: my_project
level_of_theory: wb97xd/def2tzvp
rotor_scan_resolution: 4.0
species:
  - label: ethane
    smiles: CC

Route (input → job), through the existing per-job override channel rather than a new one:

ARC.__init__(rotor_scan_resolution=...)  ->  validated, stored, round-tripped via as_dict()
ARC.execute()  ->  Scheduler(rotor_scan_resolution=...)
run_job()      ->  set_scan_resolution(args, job_type)  ->  args['trsh']['scan_res'] = value
job_factory -> JobAdapter -> set_job_attributes  ->  scan_res used by the ESS scan

• Only job_type == 'scan' (the continuous 1D ESS scans) is affected. Brute-force ND directed
  scans (job_type='directed_scan', which use dihedral_increment) and the pipe path are untouched.
• An explicit troubleshooting scan_res is never overridden — troubleshooting still wins.

Guard: refuse a too-coarse value

A resolution coarser than 20° (fewer than 18 points per rotor) is refused with InputError at
input-parse time, not merely warned. Below 18 points RMG-Py's Fourier fitter never runs and
get_potential() reads an uninitialised C double — a coarse value is not a preference but a silent
wrong answer, which is exactly the incident above. 20.0° (18 points) is the coarsest still accepted.

Backwards compatibility

Absent the key, behaviour is byte-identical to before: rotor_scan_resolution defaults to
None, as_dict() omits it (so restart.yml is unchanged), and set_scan_resolution leaves the
job args untouched so the settings default (8.0) still reaches the job.

Tests

Four new tests (parse/store/round-trip; absent-key no-op; the guard; and a real scan JobAdapter
built via job_factory receiving the value and falling back to the settings default when unset).

• arc/main_test.py arc/scheduler_test.py: 63 passed.
• arc/job/adapters/common_test.py arc/job/adapters/gaussian_test.py: 31 passed.

Run under arc_env (Python 3.14) with HOME pointed at an empty dir (ARC's suite reads
~/.arc/settings.py) and -n0 -o addopts="".

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 64.59%. Comparing base (a08d314) to head (f6ceb91).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1019      +/-   ##
==========================================
+ Coverage   64.55%   64.59%   +0.04%     
==========================================
  Files         119      119              
  Lines       39788    39805      +17     
  Branches    10307    10312       +5     
==========================================
+ Hits        25684    25711      +27     
+ Misses      11123    11117       -6     
+ Partials     2981     2977       -4     

Flag	Coverage Δ
functionaltests	64.59% <ø> (+0.04%)	⬆️
unittests	64.59% <ø> (+0.04%)	⬆️

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.

[Copilot]

Pull request overview

Adds a run-level rotor_scan_resolution input key so 1D rotor scan resolution is owned by the run (not host settings), validated at input parse time, and routed through the existing per-job troubleshooting override channel to downstream scan jobs.

Changes:

• Introduce rotor_scan_resolution on ARC with validation + restart round-tripping via as_dict().
• Thread the value into Scheduler and inject it into scan jobs via args['trsh']['scan_res'] when appropriate.
• Add unit tests covering parsing/round-tripping, guard behavior, and propagation into real scan jobs.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
arc/main.py	Adds rotor_scan_resolution to ARC, validates it, and persists it in restart dicts when set.
arc/scheduler.py	Accepts run-level rotor_scan_resolution and injects it into scan job args without overriding explicit troubleshooting.
arc/main_test.py	Adds tests for parsing/round-tripping and validation guard behavior.
arc/scheduler_test.py	Adds tests ensuring injection happens only for scan jobs and reaches JobAdapter.scan_res.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.