convert_salary_excel.py detected the company column via exact membership
in COMPANY_PATTERNS, so common real-world headers like "Company Name" or
"Employer Name" were never matched. parse_sheet then returned [] for that
sheet, silently dropping it from salary_data.json (or exiting with no
output for a single-sheet file).
Route company-column detection through the existing header_matches()
token matcher (already used for count/index detection). This only adds
detections; inputs that already worked (bare "Company"/"Firma"/...) are
unaffected.
Adds a regression test in tests/test_convert_salary_excel.py that fails
on master (returns []) and passes after the fix.
This commit improves the performance and consistency of the salary tools:
- Redundant query normalization and word extraction are eliminated in salary_lookup.py by pre-calculating representations once before the search loop.
- A match_score_optimized helper is introduced to perform the comparison using the pre-calculated query data, preserving full backward compatibility for match_score.
- Normalization in tools/convert_salary_excel.py is unified: paired column headers now consistently substitute spaces and dashes with underscores (e.g. 'software_engineering') to match the single-column formatting.
- Unit test coverage is significantly expanded in tests/test_salary_lookup.py and tests/test_convert_salary_excel.py to cover normalization, anglicization, search filtering, and matching behaviors.
* refactor(salary): make compound-word matching locale-agnostic
The Excel column detector hardcoded a DANISH_COMPOUND_PATTERNS set inside
header_matches(), so the compound-word matching that helps Danish headers
(e.g. "lønindeks") was baked into the algorithm by name and unavailable to
any other locale without editing the source.
Rename it to COMPOUND_PATTERNS and pass it as a parameter (default
unchanged, so the Danish demonstration data behaves identically). A
different-locale spreadsheet can now supply its own compound tokens via
header_matches(..., compound_patterns=...). Add a test covering both the
preserved default and the parameterized path.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* refactor(salary): drop unused compound_patterns parameter
Per review: keep the DANISH_COMPOUND_PATTERNS -> COMPOUND_PATTERNS
rename (universal template naming, defaults still Danish), but remove
the compound_patterns= parameter. No caller passes a custom set, and a
fork adapting another locale edits the module-level constant either way,
so parameterizing it is speculative generality per CONTRIBUTING.md.
header_matches() now reads COMPOUND_PATTERNS directly. Test updated to
verify compound-vs-whole-token matching against the constant.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Alaa-Taieb