potential/backend: coerce scalar coords for backend-compatible scalar-only methods

[jobovy]

What

Coerce scalar coordinates onto the active backend inside check_potential_inputs_not_arrays, for _backend_compatible potentials only.

The scalar-only compute methods of migrated potentials (RotateAndTiltWrapperPotential, DoubleExponentialDiskPotential) still do real backend arithmetic on their coordinates — xp.isinf(R), xp.cos(phi), xp.sin(phi) — which torch rejects on a plain python float under a forced backend:

TypeError: isinf(): argument 'input' (position 1) must be Tensor, not float
   at RotateAndTiltWrapperPotential.py:174  (Rinf = xp.isinf(R))

This was the root cause of the mockRotatedAndTilted* test_amp_mult_divide torch failures (the wrapper's normalize() and _evaluate reach the scalar-only methods with python-float R/z/phi).

 if (... array-shape check ...):
     raise TypeError(...)
+if getattr(self, "_backend_compatible", False):
+    xp = get_namespace(R, z, phi)
+    R, z, phi = coerce_coords(xp, R, z, phi)
 return func(self, R, z, phi, t)

The gate (caught by adversarial review)

An earlier unconditional version of this coercion regressed the unmigrated AnyAxisymmetricRazorThinDiskPotential — its bare scipy.integrate.quad / numpy internals reject a 0-d tensor ('<' not supported between numpy.ndarray and Tensor), turning ~9 currently-green forced-torch tests red. The fix mirrors the existing potential_physical_input gate (conversion.py: if xp is not numpy and _check_backend_compatible(Pot)): only coerce for _backend_compatible potentials, so unmigrated scalar-only potentials keep their python-float inputs. t is left uncoerced (it may be a hashable cache key downstream).

Result

• Flips the RotateAndTilt-wrapped test_amp_mult_divide torch entries: mockRotatedAndTiltedMWP14WrapperPotential[wInclination] standalone; the TriaxialLogHalo-wrapped one composes with potential/backend: centrally coerce forced-backend scalar leaks #987's phi-default coercion (all 4 pass on feat/backends+potential/backend: centrally coerce forced-backend scalar leaks #987).
• numpy byte-identical: coerce_coords is an object-identical pass-through when xp is numpy, and the gate skips it entirely for unmigrated potentials.
• No regression: test_backend_disk.py+test_backend_diskexpansion.py (DoubleExp) 331✓, test_backend_wrappers.py 480✓, AnyAxiRazorThin forced-torch restored to baseline (the one ledgered test_normalize_potential xfail unchanged).
• Ledger untouched (flipped entries become XPASS, green via strict=False; pruned in the separate ledger-regen PR).

Tests

• test_scalar_only_python_float_inputs_forced_backend (test_backend_wrappers.py): passes plain python floats to RotateAndTilt's _evaluate/_Rforce/_zforce/_phitorque/_dens under a forced backend, so the decorator coercion is what's under test. Fails at the isinf line without the fix.
• test_scalar_only_gate_spares_unmigrated_potential (test_backend_conventions.py): asserts AnyAxisymmetricRazorThinDiskPotential (not _backend_compatible) still evaluates under a forced backend — the regression guard for the gate.

The new coercion lines run on the numpy path too (pass-through), so they are covered by existing numpy tests.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.93%. Comparing base (e0d454d) to head (a64daf6).
⚠️ Report is 1 commits behind head on feat/backends.

Additional details and impacted files

@@              Coverage Diff               @@
##           feat/backends     #990   +/-   ##
==============================================
  Coverage          99.93%   99.93%           
==============================================
  Files                254      254           
  Lines              39833    39836    +3     
  Branches             834      840    +6     
==============================================
+ Hits               39808    39811    +3     
  Misses                25       25           

☔ 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.

[github-actions]

All-backend test status (jax / torch)

Commit f6d832fd6aaca28a1c2024d33c7258fbfa6f127d

Green is achieved via the checked-in xfail-ledger (tests/backend_xfail.txt, applied xfail(strict=False)), so the metric to watch is the shrinking xfail count (burndown), not a raw pass count. A FAIL/ERR is an un-ledgered regression (reds the run). Because the ledger is non-strict, a now-passing ledgered test is a plain pass here (no per-push XPASS); burndown candidates -- in both directions -- are surfaced by the scheduled regen run, which rewrites the ledger from real outcomes. deferred is a separate burndown: tests skipped because they are unrunnable under the backend until the port is vectorized (see tests/backend_slow_skip.txt), e.g. the jax spherical-DF sampling/quadrature tests pending the Track F DF migration.

Overall: jax: 1057 passed · 265 xfail · 725 deferred | torch: 883 passed · 1160 xfail · 1 deferred

Ledger size: 2357 entries (jax=284, torch=2073).

Test shard	jax	torch
actionAngle	✅ 112 pass · 89 xfail	✅ 37 pass · 164 xfail
sphericaldf	✅ 164 pass · 26 xfail · 28 deferred	✅ 8 pass · 210 xfail
conversion + util + misc	✅ 86 pass · 5 xfail · 1 deferred	✅ 42 pass · 50 xfail
potential + scf + multipole	— (no result)	— (no result)
quantity + coords	✅ 287 pass · 49 xfail	✅ 206 pass · 130 xfail
orbit (energy/Jacobi + from_name)	✅ 0 pass · 0 xfail · 115 deferred	✅ 66 pass · 49 xfail
orbit + orbits (main)	✅ 0 pass · 0 xfail · 578 deferred	✅ 265 pass · 310 xfail
evolveddiskdf	✅ 35 pass · 0 xfail	✅ 32 pass · 3 xfail
jeans + dynamfric	✅ 17 pass · 2 xfail · 2 deferred	✅ 7 pass · 13 xfail · 1 deferred
qdf + pv2qdf + streamgapdf_impulse + noninertial	✅ 57 pass · 75 xfail · 1 deferred	✅ 14 pass · 119 xfail
streamgapdf	✅ 28 pass · 2 xfail	✅ 28 pass · 2 xfail
diskdf	✅ 129 pass · 0 xfail	✅ 112 pass · 17 xfail
streamdf + streamspraydf + streamTrack	✅ 142 pass · 17 xfail	✅ 66 pass · 93 xfail

Per-shard counts

Test shard	backend	pass	xfail	deferred	XPASS	fail	error
actionAngle	jax	112	89	0	0	0	0
actionAngle	torch	37	164	0	0	0	0
sphericaldf	jax	164	26	28	0	0	0
sphericaldf	torch	8	210	0	0	0	0
conversion + util + misc	jax	86	5	1	0	0	0
conversion + util + misc	torch	42	50	0	0	0	0
potential + scf + multipole	jax	—	—	—	—	—	—
potential + scf + multipole	torch	—	—	—	—	—	—
quantity + coords	jax	287	49	0	0	0	0
quantity + coords	torch	206	130	0	0	0	0
orbit (energy/Jacobi + from_name)	jax	0	0	115	0	0	0
orbit (energy/Jacobi + from_name)	torch	66	49	0	0	0	0
orbit + orbits (main)	jax	0	0	578	0	0	0
orbit + orbits (main)	torch	265	310	0	0	0	0
evolveddiskdf	jax	35	0	0	0	0	0
evolveddiskdf	torch	32	3	0	0	0	0
jeans + dynamfric	jax	17	2	2	0	0	0
jeans + dynamfric	torch	7	13	1	0	0	0
qdf + pv2qdf + streamgapdf_impulse + noninertial	jax	57	75	1	0	0	0
qdf + pv2qdf + streamgapdf_impulse + noninertial	torch	14	119	0	0	0	0
streamgapdf	jax	28	2	0	0	0	0
streamgapdf	torch	28	2	0	0	0	0
diskdf	jax	129	0	0	0	0	0
diskdf	torch	112	17	0	0	0	0
streamdf + streamspraydf + streamTrack	jax	142	17	0	0	0	0
streamdf + streamspraydf + streamTrack	torch	66	93	0	0	0	0