commodutil 5.5.1__tar.gz → 5.5.3__tar.gz

{commodutil-5.5.1 → commodutil-5.5.3}/CHANGELOG.md

@@ -1,5 +1,13 @@
 # Changelog
 
+## 5.5.3 - 2026-07-28
+
+- Fixed country-name resolution for standalone `St`/`St.` abbreviations and the vendor spellings `Taiwan, China` and `Virgin Islands (US)`.
+
+## 5.5.2 - 2026-07-28
+
+- Fixed country-name resolution to normalize ampersands to `and`, matching ISO names such as Antigua and Barbuda.
+
 ## 5.5.0 - 2026-07-27
 
 - Added `commodutil.standards.geography` with pycountry-backed countries, desk aliases, and membership-defined country groups; added `pycountry` as a production dependency.

{commodutil-5.5.1 → commodutil-5.5.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: commodutil
-Version: 5.5.1
+Version: 5.5.3
 Summary: common commodity/oil analytics utils
 Author-email: aeorxc <author@example.com>
 License-Expression: MIT

{commodutil-5.5.1 → commodutil-5.5.3}/commodutil/standards/geography.py

@@ -15,6 +15,7 @@ memberships are derived from those definitions at import time.
 
 from __future__ import annotations
 
+import re
 from dataclasses import dataclass
 from typing import Dict, Optional, Tuple
 
@@ -69,11 +70,13 @@ COUNTRY_ALIASES: dict[str, str] = {
     "Turkey": "TR",
     "Turkiye": "TR",
     "Republic of China": "TW",
+    "Taiwan, China": "TW",
     "USA": "US",
     "U.S.": "US",
     "U.S.A.": "US",
     "Holy See": "VA",
     "US Virgin Islands": "VI",
+    "Virgin Islands (US)": "VI",
     # From empirical vendor-vocabulary validation (see commodflows research).
     # ISO merges Bonaire and Sint Eustatius into BQ; sub-island detail is
     # lost by design.
@@ -98,6 +101,13 @@ COUNTRY_ALIASES: dict[str, str] = {
 }
 
 
+def _normalize_country_name(name: str) -> str:
+    """Return the case-folded lookup key for a country name."""
+    normalized = name.replace("&", " and ")
+    normalized = re.sub(r"\bSt\.?(?=\s|$)", "Saint", normalized, flags=re.IGNORECASE)
+    return " ".join(normalized.split()).casefold()
+
+
 def _build_country_name_map(aliases: Dict[str, str]) -> Dict[str, str]:
     """Build the case-folded name map and reject conflicting aliases."""
     result: Dict[str, str] = {}
@@ -106,7 +116,7 @@ def _build_country_name_map(aliases: Dict[str, str]) -> Dict[str, str]:
             name = getattr(country, attribute, None)
             if name is None:
                 continue
-            key = name.casefold()
+            key = _normalize_country_name(name)
             existing = result.get(key)
             if existing is not None and existing != country.alpha_2:
                 raise ValueError(
@@ -120,7 +130,7 @@ def _build_country_name_map(aliases: Dict[str, str]) -> Dict[str, str]:
             raise ValueError(
                 f"country alias {alias!r} has unknown country code {code!r}"
             )
-        key = alias.casefold()
+        key = _normalize_country_name(alias)
         existing = result.get(key)
         if existing is not None and existing != code:
             raise ValueError(
@@ -144,7 +154,7 @@ def canonical_country(text: object) -> Optional[str]:
     code = cleaned.upper()
     if code in VALID_COUNTRY_CODES:
         return code
-    return COUNTRY_NAME_MAP.get(cleaned.casefold())
+    return COUNTRY_NAME_MAP.get(_normalize_country_name(cleaned))
 
 
 def is_canonical_country(code: object) -> bool:

{commodutil-5.5.1 → commodutil-5.5.3}/commodutil.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: commodutil
-Version: 5.5.1
+Version: 5.5.3
 Summary: common commodity/oil analytics utils
 Author-email: aeorxc <author@example.com>
 License-Expression: MIT

{commodutil-5.5.1 → commodutil-5.5.3}/tests/test_standards_geography.py

@@ -66,6 +66,33 @@ def test_country_name_map_contains_pycountry_names_and_aliases():
     assert COUNTRY_NAME_MAP["uk"] == "GB"
 
 
+@pytest.mark.parametrize(
+    "value,expected",
+    [
+        ("Antigua & Barbuda", "AG"),
+        ("Trinidad & Tobago", "TT"),
+        ("Bosnia & Herzegovina", "BA"),
+        ("&", None),
+    ],
+)
+def test_canonical_country_normalizes_ampersands(value, expected):
+    assert canonical_country(value) == expected
+
+
+@pytest.mark.parametrize(
+    "value,expected",
+    [
+        ("St Kitts & Nevis", "KN"),
+        ("St. Lucia", "LC"),
+        ("Taiwan, China", "TW"),
+        ("Virgin Islands (US)", "VI"),
+        ("Stockholm", None),
+    ],
+)
+def test_canonical_country_normalizes_live_vendor_spellings(value, expected):
+    assert canonical_country(value) == expected
+
+
 @pytest.mark.parametrize(
     "value,expected",
     [

{commodutil-5.5.1 → commodutil-5.5.3}/uv.lock

@@ -1627,7 +1627,7 @@ wheels = [
 
 [[package]]
 name = "twine"
-version = "6.2.0"
+version = "7.0.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "id" },
@@ -1640,9 +1640,9 @@ dependencies = [
     { name = "rich" },
     { name = "urllib3" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/e0/a8/949edebe3a82774c1ec34f637f5dd82d1cf22c25e963b7d63771083bbee5/twine-6.2.0.tar.gz", hash = "sha256:e5ed0d2fd70c9959770dce51c8f39c8945c574e18173a7b81802dab51b4b75cf", size = 172262, upload-time = "2025-09-04T15:43:17.255Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/92/3c/58f808a359700f39a967dffede33efeac809262c03303fa3eec6afff8f49/twine-7.0.0.tar.gz", hash = "sha256:85cdb29c518efef867360ae4acd4b0dfd61c8654a22fca08e6f8539f05022177", size = 215032, upload-time = "2026-07-27T15:59:00.825Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/3a/7a/882d99539b19b1490cac5d77c67338d126e4122c8276bf640e411650c830/twine-6.2.0-py3-none-any.whl", hash = "sha256:418ebf08ccda9a8caaebe414433b0ba5e25eb5e4a927667122fbe8f829f985d8", size = 42727, upload-time = "2025-09-04T15:43:15.994Z" },
+    { url = "https://files.pythonhosted.org/packages/96/08/ddcdc06225eaad6de0e48e1002b06d919dbde20582d0662c7af51308e5d6/twine-7.0.0-py3-none-any.whl", hash = "sha256:b854164df26db268af05f49aa5c0344b10e27a494343ff05b1e0bad3b135f5a7", size = 43204, upload-time = "2026-07-27T15:58:59.26Z" },
 ]
 
 [[package]]