Add beam-footprint utilities and PSF metrics for the wavefield path

[jacobdparker]

Summary

Stacked on #173. Promotes the patterns proven while modeling ESIS into general utilities, so a physical-optics model of a real (multi-channel, obscured) instrument is a few lines instead of a page:

• SequentialSystem.footprint(surface, ...): positions where a dense pupil grid of rays traced through the entire system (all masks active) intersects a given surface, in local surface coordinates, plus the mask of rays that reach the sensor - i.e. the used portion of each optic. Useful for geometric work independently of wavefields.
• optika.apertures.footprint_aperture(position, where): a PolygonalAperture from the convex hull of a footprint - the constructor for an effective channel pupil (e.g. the sector of the ESIS primary delimited by the obscuration hub, the octagon edge, and the projected grating trapezoid).
• wavefield(bound="footprint", padding_relative=..., padding_absolute=...): automatic per-surface sampling regions from the beam footprint. The absolute padding floor handles intermediate-focus surfaces, where the geometric footprint collapses to a point but the diffraction pattern needs a finite window.
• PSF metrics (optika.wavefields.encircled_energy_radius, ensquared_energy, fwhm): image-quality numbers that were previously recomputed by hand for every analysis; ensquared_energy is the natural pixel-limited criterion. Validated against an analytic Gaussian, including broadcasting over extra axes.
• A jupyter-execute example on SequentialSystem.psf (footprint-bounded Huygens PSF of a parabolic telescope).
• scipy promoted from transitive to explicit dependency (scipy.spatial.ConvexHull).

Validation

Reproducing the hand-built ESIS f1 channel model through this API (footprint + footprint_aperture + bound="footprint" with padding_absolute=50 um) matches the reference results: image centroid within 0.13 um of the geometric raytrace and EE50 = 3.97 um vs the 3.98 um reference.

🤖 Generated with Claude Code

[codecov]

Codecov Report

❌ Patch coverage is 97.98995% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 98.62%. Comparing base (60267ad) to head (3010b30).

Files with missing lines	Patch %	Lines
optika/wavefields/_metrics.py	95.52%	3 Missing ⚠️
optika/apertures/_apertures.py	95.83%	1 Missing ⚠️

Additional details and impacted files

@@                  Coverage Diff                   @@
##           feature/wavefields     #174      +/-   ##
======================================================
- Coverage               98.64%   98.62%   -0.02%     
======================================================
  Files                     127      129       +2     
  Lines                    6849     7046     +197     
======================================================
+ Hits                     6756     6949     +193     
- Misses                     93       97       +4     

Flag	Coverage Δ
unittests	98.62% <97.98%> (-0.02%)	⬇️

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.