Fix bugs in core classes + add test suite and CI pipeline

[kartikeyg0104]

Closes #10

Bug fixes

1. EmissionActivityRatio typo in Config.py — Fixed 'e''t' (Python string concatenation → 'et') to 'e','t' in PARAMETERS_C. This was silently corrupting parameter validation dimensions.

2. File handle leaks in FileClass.py — Refactored all four static methods (readFile, writeFile, writeFileUJson, readParamFile) from manual open()/close() to with open(...) as f: context managers. Previous code leaked file handles when exceptions occurred between open() and close().

3. CustomThread.join() missing timeout — Added timeout=None parameter to match Thread.join() signature. Previously, thread.join(timeout=5) would raise TypeError.

Testing infrastructure

4. Added pytest test suite — 25 tests across 3 test files:

   • tests/test_config.py (9 tests) — validates PARAMETERS_C dimensions, config paths, group tuples, and mapping key consistency
   • tests/test_file_class.py (10 tests) — JSON read/write round-trips, missing file handling, invalid JSON handling, overwrite behavior
   • tests/test_custom_thread.py (6 tests) — return values, args/kwargs passing, timeout acceptance, timeout behavior

5. Added GitHub Actions CI (.github/workflows/ci.yml):

   • Lint job: flake8 checking for syntax errors (E9, F63, F7, F82)
   • Test job: pytest across matrix of Python 3.10/3.11/3.12 × Ubuntu/macOS/Windows

6. Added pytest.ini — configures test discovery paths and naming conventions

7. Added __pycache__/ to .gitignore

Validation

$ pytest tests/ -v --tb=short
============================== 25 passed in 0.73s ==============================

All 25 tests pass locally on macOS with Python 3.12.

[SeaCelo]

Thank you. Please keep this as a draft fo now until we examine the issue and weather it should be upstream or here.

[SeaCelo]

I like the testing infrastructure. Can you open that as a separate issue?

[kartikeyg0104]

I like the testing infrastructure. Can you open that as a separate issue?

Hi @SeaCelo, thanks for the feedback! I've opened #14 to track the testing infrastructure separately as you suggested. Happy to keep this PR as a draft, I'll wait for your decision on whether the bug fixes should go upstream or stay here before making any further changes. Let me know if there's anything else you'd like adjusted.

[NamanmeetSingh]

Hey, great catch on the EmissionActivityRatio dimension typo! Since we are laying the groundwork for the OG-CLEWS integration, I had a quick architectural thought regarding the data pipelines when you port this:

1. Dimension Extensibility (test_config.py):

In test_all_dimension_keys_are_single_chars_or_known, the known set strictly maps to CLEWS/OSeMOSYS dimensions (r, t, f, etc.). As we integrate OG-Core, we will need to support macroeconomic dimensions (like S for age cohorts, J for income groups, I for industries). Are we planning to expand PARAMETERS_C to hold these, or will we spin up a separate PARAMETERS_OG dictionary? Either way, we might want to flag this test so it doesn't block the CI when we start mapping the coupled-model schemas.

2. Large Dataset Memory Handling (FileClass.py):

Love the move to context managers! One quick optimization for the model runs: using json.loads(f.read()) pulls the entire text into memory before parsing, which can cause RAM spikes when loading massive multi-country scenario files. Switching to json.load(f) allows the parser to stream directly from the file, which is much safer for our heavy data payloads.

Really clean refactor overall! The file handle leak fix is definitely going to save us some debugging headaches during parallel runs.