docs: SUAVE 2.5.2 install/integration notes (feasibility + API validation)

[aligatoralex]

Description of Changes:

Adds docs/SUAVE_INTEGRATION_NOTES.md — a hands-on report from actually installing SUAVE 2.5.2 (the pinned external/suave submodule) into an isolated venv, kept entirely out of the project's uv environment (pyproject.toml / uv.lock untouched, submodule untouched). Created in response to the question "why not just install SUAVE?" — the answer, with evidence.

Key findings:

• It installs, but not cleanly. pip install -e needs --no-build-isolation (SUAVE's setup.py eagerly import numpys at build-requirement time), and import SUAVE then needs three runtime shims applied first — all on SUAVE's eager import chain, so unavoidable:
  1. collections.MutableMapping alias (SUAVE's vendored pint, broken on Python ≥3.10 — every interpreter on this box).
  2. scipy.integrate.cumtrapz → cumulative_trapezoid (a genuine drop-in).
  3. scipy.misc.derivative — removed with no drop-in; a central-difference stand-in that unblocks the import but is not numerically validated.
• VehicleFactory ([FEATURE] Update VehicleFactory to support vehicle-agnostic ComponentStore architecture #29) API validated against the real package — zero mismatches. Turbojet_Super, Ramjet, Liquid_Rocket, Scramjet all exist; Wing.sweeps/spans/taper/aspect_ratio/dihedral/vertical all present with the expected shapes. A solid-motor network is confirmed genuinely absent, which validates the translator's NotImplementedError for SolidMotor.
• The [FEATURE] Update VehicleFactory to support vehicle-agnostic ComponentStore architecture #29 SUAVE-gated integration test passes on the real install (3/3) — inspected via git archive from feature/vehicle-factory-composition without ever switching the repo branch.
• Recommendation: keep SUAVE in a dedicated venv/container behind a small, maintained suave_compat shim (turning this exploration's monkeypatches into tested code), never uv add'd into pyproject.toml; plus the CI implications for the gated test.

This validates PR #30 (VehicleFactory) end-to-end against a real SUAVE and turns "why not install SUAVE" into a documented dependency-isolation decision.

Checklist:

• I have run uv run pytest YAADO_Core/tests/ and all tests pass. (Docs-only; the project uv env was deliberately left untouched — SUAVE went into a throwaway venv.)
• My code includes type hints and docstrings. (N/A — Markdown only.)
• I added all vital READMEs. (Adds integration notes under docs/.)

Verification method:
Every claim in the report comes from commands actually run in the isolated venv (exact error messages and the successful import/test output are reproduced in the doc). The API surface was probed directly against the installed SUAVE 2.5.2; the gated #29 test was executed against it (3 passed).

---

Generated by Claude Code

[aligatoralex]

Superseded by #47 — consolidated all docs-only RFC/notes PRs into a single docs/rfcs/ + docs/research/ PR (see #47 for rationale: avoiding one-branch-per-document sprawl, matching Rust RFCs/Python PEPs/Kubernetes KEPs/ADR conventions). docs/SUAVE_INTEGRATION_NOTES.md moves in unchanged in #47. Closing this PR; its branch will be deleted.

---

Generated by Claude Code