chore: 0.5.0 clean-up -- move TerraME validation to disslucc-benchmark

[profsergiocosta]

Description

Clean-up release (0.5.0). The numerical comparison with TerraME/LuccME now lives in
LambdaGeo/disslucc-benchmark, which
pins a disslucc release; this repository keeps the package and examples to learn it
with. Everything removed is available in the v0.4.0 tag, which is untouched (the
dissmodel JOSS paper cites it).

• Removed benchmark/goldens/ and the golden, per-year, discriminance, validation and
  Lua differential tests with their helpers; dropped the lupa dev dependency and
  examples/run_lab1_validation.py / run_lab15_validation.py.
• Renamed examples without "lab": run_continuous_real_data.py,
  run_continuous_executor.py, run_discrete_executor.py; TOML headers now say how to
  run each file locally with the CLI.
• Added tests/test_executor_toml.py: each example TOML, run through the CLI, writes the
  same output file (same SHA-256) as the equivalent hand-built experiment (continuous),
  and runs reproducibly (discrete, saturation).
• docs/validation.md is now a short pointer to disslucc-benchmark and v0.4.0;
  README, quickstart, index, CLAUDE.md, CONTRIBUTING and issue templates updated.
• Fixed 4 pre-existing mypy errors in components/demand/dates.py (two
  # type: ignore[override], one int→float conversion).
• Added .github/workflows/publish.yml (PyPI Trusted Publishing on release).
• Version 0.5.0; CHANGELOG.md and CITATION.cff updated (repository URL unified to
  DisSModel/disslucc, author name order as in dissmodel).

No change to src/ behaviour. disslucc-benchmark's 22 tests pass against this branch.

Related Issue

Closes # n/a

Changes Made

• Core implementation / algorithmic changes (src/disslucc/components/, executors/)
• Added or updated unit tests (tests/)
• Updated documentation / docstrings / examples

Contributor Checklist

• The code runs locally without errors
• All unit tests pass locally (pytest tests/ -v) — 23 passed
• Static analysis passes (mypy src/disslucc) — also ruff check .
• Docstrings follow the project standard (inputs, outputs, exceptions) — no new public API
• If this changes a Lab1 or Lab15 result, docs/validation.md (and
  paper.md/paper.bib in dissmodel, if applicable) is updated to
  match -- no result changes; docs/validation.md now points to disslucc-benchmark,
  and the dissmodel paper (separate PR) cites the benchmark ids
  (lab01_md1643, lab15_md10)

After merge

Release v0.5.0 on GitHub (publishes to PyPI and creates the Zenodo DOI), then pin
disslucc==0.5.0 in disslucc-benchmark and tag it v0.2.0.