Skip to content

implementation_plan.md

Latest commit

 

History

History
22 lines (16 loc) · 1.67 KB

implementation_plan.md

File metadata and controls

22 lines (16 loc) · 1.67 KB

Fixing Failing Tests Implementation Plan

1. Goal Description

The recent changes to LANGUAGE_STANDARDS and the ExcelReader class have broken some existing tests in test_extract.py and test_language_parser.py. The goal is to update these tests to match the new desired behaviors.

  1. test_extract.py: The ExcelReader was refactored significantly to remove global forward filling (ffill()) for merged cells. That responsibility was intentionally moved to DataCleaner. Therefore, test_read_merged_cells is no longer a valid test for ExcelReader and will be updated to expect NaN for merged cells, or removed.
  2. test_language_parser.py: The regional codes for Chinese and Japanese changed from B1~B3 to CN_B1~CN_B3 and C1~C2 to JP_C1~JP_C2. The assertions in test_chinese_codes_exist, test_japanese_codes_exist, and test_parse_b1_chinese still expect the old keys. These will be updated.

2. Proposed Changes

2.1 Update tests/test_extract.py

  • Modify test_read_merged_cells so it correctly asserts that the second row's nation value for a merged cell is nan (or technically None / float('nan')), demonstrating that ExcelReader is no longer incorrectly forward-filling data, which is now handled downstream.

2.2 Update tests/test_language_parser.py

  • Modify test_chinese_codes_exist to loop over ["CN_B1", "CN_B2", "CN_B3"].
  • Modify test_japanese_codes_exist to loop over ["JP_C1", "JP_C2"].
  • Modify test_parse_b1_chinese to assert level_code == "CN_B1" instead of "B1".

3. Verification Plan

  • Automated Tests:
    • Run python -m pytest tests/ -v
    • All tests should pass without any AssertionError.