Route package diagnostics through the logging module

[thc1006]

Route the package's bare print() diagnostics through the standard logging module so verbosity is controllable: quiet during scoring or CI, verbose when debugging. Implements the logging instrumentation thread of #14.

What changed

• balloon_world.py and evaluate.py each get a module logger (logging.getLogger(__name__)).
• balloon_world.py: the two env termination messages become logger.info, and the close() trace becomes logger.debug.
• evaluate.py: evaluate_scenario's two result lines become logger.info (lazy %s formatting). The function now also returns info["popped_count"], so library callers get the result, since the diagnostics are silent unless the caller configures logging. The __main__ entry point calls logging.basicConfig(level=INFO, stream=sys.stdout) so running python evaluate.py <cfg> still prints results to stdout.
• The two example callers, doc/examples/run_env_agent.py and the Colab notebook, now display the returned reward.
• New tests/test_logging.py: an AST invariant that fails if a bare print() reappears in the package.
• New tests/test_evaluate.py: pins evaluate_scenario's return contract.

Behaviour

Reward computation and env logic are unchanged. Env diagnostics are silent by default (the consumer opts in via logging); evaluate_scenario's result is available both as a return value and, on the script and configured-logging paths, as log output.

The logger uses getLogger(__name__), so a consumer can control the whole package with logging.getLogger("BalloonPoppingGymEnv"). This minimal, standard setup composes with any later logging integration.

About #14

#14 covered three areas: functional test coverage, logging instrumentation, and a coverage CI workflow. CI and coverage are handled by #21 (PR #23); logging is this PR.

Verified

• TDD: test_logging.py started red (5 print() offenders) and went green after the change. test_evaluate.py started red (evaluate_scenario returned None) and went green after.
• ruff check and ruff format --check pass; the full test suite passes (11 tests).
• The Colab evaluation path was run end to end in a fresh submodule-based environment: evaluate_scenario returns a real count and the notebook prints Total reward.

Closes #14

[Copilot]

Pull request overview

Routes the package's bare print() diagnostics through the standard logging module so verbosity is controllable (quiet in scoring/CI, verbose when debugging), and adds an AST-based test to prevent regressions.

Changes:

• Add module loggers in balloon_world.py and evaluate.py; replace print() calls with logger.info / logger.debug.
• Configure logging.basicConfig(level=INFO) in evaluate.py's __main__ block to preserve CLI output.
• Add tests/test_logging.py AST invariant that fails if any bare print() reappears in the package.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
BalloonPoppingGymEnv/envs/balloon_world.py	Replace three print() diagnostics with module logger calls (info/debug).
BalloonPoppingGymEnv/evaluation/evaluate.py	Replace two CLI prints with lazy-formatted logger.info; add basicConfig only in __main__.
tests/test_logging.py	New AST test enforcing no bare print() in the package.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.