Honor digit_group_separator's complementary decimal separator

[binggao1230]

Summary

Price.fromstring documents that digit_group_separator implies the complementary decimal separator:

If digit_group_separator is ".", then 1.000 is parsed as 1000. If it is ",", then 1.000 is parsed as 1.

The "," case is broken:

>>> from price_parser import Price
>>> Price.fromstring("1.000", digit_group_separator=",").amount
Decimal('1000')        # docstring says this should be 1 (i.e. Decimal('1.000'))

fromstring strips the group separator but leaves decimal_separator=None, so parse_number re-guesses the remaining ./, — and for "1.000" it guesses the . is a thousands group. The sibling paths confirm the inconsistency: decimal_separator="." on the same input correctly gives Decimal('1.000'), and digit_group_separator="." correctly gives 1000 — only the ","-group branch was wrong.

Fix

When the caller declares one of "."/"," as the digit group separator and passes no explicit decimal_separator, treat the other symbol as the decimal separator — exactly the documented semantics:

if decimal_separator is None and digit_group_separator in (".", ","):
    decimal_separator = "," if digit_group_separator == "." else "."

Non-./, group separators (e.g. a space) and the explicit-decimal_separator path are unaffected.

Verification

• Reproduced on master (64e213a): all four docstring claims now hold, and mixed inputs like "1,000.50" (group="," → 1000.50) and "1.000,50" (group="." → 1000.50) parse correctly.
• Added a parametrized test_price_digit_group_separator mirroring test_price_decimal_separator; the documented ","-group case fails before the fix and passes after.
• Full suite: 1065 passed, 134 xfailed (baseline 1059 passed + 6 new params), no regressions. ruff and mypy clean.

---

This pull request was prepared with the assistance of AI, under my direction and review.

[AdrianAtZyte]

Could you remove this comment and instead update the docstring to reflect the new logic?

[wRAR]

(That they won't is the main reason I close such PRs)

[AdrianAtZyte]

🤞

[binggao1230]

Done — comment dropped, docstring now covers the inference. Sorry for the long delay.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (64e213a) to head (1e6e740).

Additional details and impacted files

@@             Coverage Diff             @@
##           master       #77      +/-   ##
===========================================
+ Coverage   98.23%   100.00%   +1.76%     
===========================================
  Files           3         3              
  Lines         113       115       +2     
  Branches       16        17       +1     
===========================================
+ Hits          111       115       +4     
+ Misses          1         0       -1     
+ Partials        1         0       -1     

Files with missing lines	Coverage Δ
price_parser/parser.py	100.00% <100.00%> (+2.32%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[AdrianAtZyte]

Do you think you could complete test coverage based on the feedback from CodeCov?

[binggao1230]

Added a test for the explicit-decimal-separator case (1e6e740) — the group-separator inference now has full branch coverage (patch 100%).