Add scripts for OpenAPI generation and post-generation and more convenience

[antalszava]

Summary

• Adds scripts/regenerate_models.py as the OS-independent way to regenerate the client
• Replace the perl one-liner post_hooks in openapi-python-client-config.yaml with scripts/post_generate.py (same behavior: SPDX/@generated headers, token hidden from repr) — the perl quoting was Unix-only and broke generation on Windows.
• Add an optional Makefile with shorthand targets (regen, sync-spec, check-generated, lint, test, ...) that wrap the script and the existing uv run commands.
• generated.yml now runs the script on a Linux/macOS/Windows matrix (staleness gate on Linux), so the regeneration command is defined in one place and exercised on all supported platforms.
• ci.yml test matrix gains macOS and Windows entries (Python 3.14, --no-cov).
• Update CONTRIBUTING.md to document the script as the regeneration workflow.

Test plan

• uv run --group regen python scripts/regenerate_models.py reproduces the committed generated code with zero diff (i.e. the Python post-hooks are byte-identical to the perl ones).
• Full suite passes locally: 242 tests, 100% branch coverage on hand-written code; ruff check, ruff format --check, and ty check clean.
• generated.yml staleness jobs pass on ubuntu, macos, and windows on this PR.

Note: Used the help of Claude Code for this PR. 🤖

[splch]

Nice cleanup. One Windows bug inline + two notes. Per plan on #95, this lands after the stack.

[hodgestar-ionq]

Looks good to me. Thank you for adding the Makefile. I left a couple of questions.

[splch]

Heads-up: #101 added spec-sync.yml, which carries the same regeneration invocation as generated.yml. If this PR changes how regeneration is invoked, please update spec-sync.yml in the same change so the weekly bot regen and the staleness check stay identical.

[splch]

zizmor noo!