main.py runs at import time, so its render loop cannot be unit tested

[dmccoystephenson]

Observation

main.py runs work at import time. Its argument parsing, its LocationService /
EnvironmentService construction and the bare main() call at the bottom of the file
all execute as soon as the module is imported:

locationService = LocationService(url, port)
environmentService = EnvironmentService(url, port)
...
main()

There is no if __name__ == "__main__": guard.

Consequence

The module cannot be imported by a test without the whole program running, so nothing
in main.py can be covered by the tests/ suite introduced in PR #10 — including the
RenderWindow-driven render loop that PR refactors, the environments.json cache-key
logic, and the --exit-after-create early-return path. The refactor in PR #10 was
therefore validated only by a hand-built headless harness that stubbed the Viron
services and injected the module chain into sys.modules, which is not something a
reviewer can re-run from the repository.

Suggested change

• Guard the entry point with if __name__ == "__main__": so importing the module has
  no side effects. python main.py is unaffected, since the module is __main__ in
  that case, and create_environments.bat invokes it the same way.
• Move the module-level argument parsing and service construction into a function
  (for example parse_args() and a main(gridSize, exitAfterCreate, ...) signature),
  so a test can drive main() with mocked services.
• Add tests/test_main.py covering, at minimum, the grid-size fallback on unparseable
  input, the environments.json cache key, and loop termination on QUIT.

Scope note

This is a refactor of the entry point rather than a behavior change, and it touches the
live render loop, so the project's live-Viron manual-validation anchor is expected to be
required for it.

This issue body was drafted during a Gardener session (https://github.com/Stephenson-Software/gardener).

---

drafted by Claude on behalf of Daniel Stephenson