hdx-python-country 3.7.3__tar.gz → 3.7.5__tar.gz

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: hdx-python-country
-Version: 3.7.3
+Version: 3.7.5
 Summary: HDX Python country code and exchange rate (fx) utilities
 Project-URL: Homepage, https://github.com/OCHA-DAP/hdx-python-country
 Author: Michael Rans
@@ -25,9 +25,9 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Requires-Python: >=3.8
-Requires-Dist: hdx-python-utilities>=3.7.0
+Requires-Dist: hdx-python-utilities>=3.7.2
 Requires-Dist: libhxl>=5.2.1
-Requires-Dist: rapidfuzz
+Requires-Dist: pyphonetics
 Provides-Extra: dev
 Requires-Dist: pre-commit; extra == 'dev'
 Provides-Extra: test

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/documentation/main.md

@@ -42,6 +42,9 @@ The code for the library is [here](https://github.com/OCHA-DAP/hdx-python-countr
 The library has detailed API documentation which can be found in the menu at the top.
 
 ## Breaking Changes
+From 3.7.5, removed clean_name function. There is now a function normalise in
+HDX Python Utilities.
+
 From 3.5.5, after creating an Adminlevel, call either setup_from_admin_info,
 setup_from_libhxl_dataset or setup_from_url.
 

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/pyproject.toml

@@ -34,9 +34,9 @@ classifiers = [
 requires-python = ">=3.8"
 
 dependencies = [
-    "hdx-python-utilities>=3.7.0",
+    "hdx-python-utilities>=3.7.2",
     "libhxl>=5.2.1",
-    "rapidfuzz",
+    "pyphonetics",
 ]
 dynamic = ["version"]
 

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/requirements.txt

@@ -8,7 +8,7 @@ attrs==23.2.0
     #   jsonlines
     #   jsonschema
     #   referencing
-certifi==2024.6.2
+certifi==2024.7.4
     # via requests
 cfgv==3.4.0
     # via pre-commit
@@ -18,17 +18,17 @@ charset-normalizer==3.3.2
     # via requests
 click==8.1.7
     # via typer
-coverage==7.5.3
+coverage==7.5.4
     # via pytest-cov
 distlib==0.3.8
     # via virtualenv
 et-xmlfile==1.1.0
     # via openpyxl
-filelock==3.15.3
+filelock==3.15.4
     # via virtualenv
 frictionless==5.17.0
     # via hdx-python-utilities
-hdx-python-utilities==3.7.0
+hdx-python-utilities==3.7.2
     # via hdx-python-country (pyproject.toml)
 humanize==4.9.0
     # via frictionless
@@ -60,7 +60,7 @@ loguru==0.7.2
     # via hdx-python-utilities
 markdown-it-py==3.0.0
     # via rich
-marko==2.1.1
+marko==2.1.2
     # via frictionless
 markupsafe==2.1.5
     # via jinja2
@@ -68,7 +68,7 @@ mdurl==0.1.2
     # via markdown-it-py
 nodeenv==1.9.1
     # via pre-commit
-openpyxl==3.1.4
+openpyxl==3.1.5
     # via hdx-python-utilities
 packaging==24.1
     # via pytest
@@ -84,12 +84,14 @@ ply==3.11
     #   libhxl
 pre-commit==3.7.1
     # via hdx-python-country (pyproject.toml)
-pydantic==2.7.4
+pydantic==2.8.2
     # via frictionless
-pydantic-core==2.18.4
+pydantic-core==2.20.1
     # via pydantic
 pygments==2.18.0
     # via rich
+pyphonetics==0.5.3
+    # via hdx-python-country (pyproject.toml)
 pytest==8.2.2
     # via
     #   hdx-python-country (pyproject.toml)
@@ -110,8 +112,6 @@ pyyaml==6.0.1
     #   frictionless
     #   pre-commit
     #   tableschema-to-template
-rapidfuzz==3.9.3
-    # via hdx-python-country (pyproject.toml)
 ratelimit==2.2.1
     # via hdx-python-utilities
 referencing==0.35.1
@@ -164,14 +164,16 @@ typing-extensions==4.12.2
     #   pydantic-core
     #   typer
 unidecode==1.3.8
-    # via libhxl
+    # via
+    #   libhxl
+    #   pyphonetics
 urllib3==2.2.2
     # via
     #   libhxl
     #   requests
-validators==0.28.3
+validators==0.30.0
     # via frictionless
-virtualenv==20.26.2
+virtualenv==20.26.3
     # via pre-commit
 wheel==0.43.0
     # via libhxl

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/src/hdx/location/_version.py

@@ -12,5 +12,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '3.7.3'
-__version_tuple__ = version_tuple = (3, 7, 3)
+__version__ = version = '3.7.5'
+__version_tuple__ = version_tuple = (3, 7, 5)

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/src/hdx/location/adminlevel.py

@@ -7,10 +7,9 @@ from hxl import InputOptions
 from hxl.input import HXLIOException
 
 from hdx.location.country import Country
-from hdx.location.names import clean_name
 from hdx.location.phonetics import Phonetics
 from hdx.utilities.dictandlist import dict_of_sets_add
-from hdx.utilities.text import multiple_replace
+from hdx.utilities.text import multiple_replace, normalise
 from hdx.utilities.typehint import ListTuple
 
 logger = logging.getLogger(__name__)
@@ -146,7 +145,7 @@ class AdminLevel:
         self.pcode_to_name[pcode] = adm_name
 
         name_to_pcode = self.name_to_pcode.get(countryiso3, {})
-        name_to_pcode[clean_name(adm_name)] = pcode
+        name_to_pcode[normalise(adm_name)] = pcode
         self.name_to_pcode[countryiso3] = name_to_pcode
         self.pcode_to_iso3[pcode] = countryiso3
         self.pcode_to_iso3[pcode] = countryiso3
@@ -156,7 +155,7 @@ class AdminLevel:
                 countryiso3, {}
             )
             name_to_pcode = name_parent_to_pcode.get(parent, {})
-            name_to_pcode[clean_name(adm_name)] = pcode
+            name_to_pcode[normalise(adm_name)] = pcode
             name_parent_to_pcode[parent] = name_to_pcode
             self.name_parent_to_pcode[countryiso3] = name_parent_to_pcode
             self.pcode_to_parent[pcode] = parent
@@ -598,7 +597,7 @@ class AdminLevel:
                 if logname:
                     self.errors.add((logname, countryiso3, parent))
                 return None
-        adm_name_lookup = clean_name(name)
+        adm_name_lookup = normalise(name)
         adm_name_lookup2 = multiple_replace(
             adm_name_lookup,
             self.get_admin_name_replacements(countryiso3, parent),

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/src/hdx/location/phonetics.py

@@ -1,18 +1,18 @@
 from typing import Callable, Optional
 
-from rapidfuzz import fuzz
+import pyphonetics
 
 from hdx.utilities.typehint import ListTuple
 
 
-class Phonetics:
+class Phonetics(pyphonetics.RefinedSoundex):
     def match(
         self,
         possible_names: ListTuple,
         name: str,
         alternative_name: Optional[str] = None,
         transform_possible_names: ListTuple[Callable] = [],
-        threshold: float = 60,
+        threshold: int = 2,
     ) -> Optional[int]:
         """
         Match name to one of the given possible names. Returns None if no match
@@ -23,22 +23,22 @@ class Phonetics:
             name (str): Name to match
             alternative_name (str): Alternative name to match. Defaults to None.
             transform_possible_names (ListTuple[Callable]): Functions to transform possible names.
-            threshold (float): Match threshold. Value is 0-100. Defaults to 60.
+            threshold: Match threshold. Defaults to 2.
 
         Returns:
             Optional[int]: Index of matching name from possible names or None
         """
-        max_similarity = 0
+        mindistance = None
         matching_index = None
 
         transform_possible_names.insert(0, lambda x: x)
 
         def check_name(name, possible_name):
-            nonlocal max_similarity, matching_index  # noqa: E999
+            nonlocal mindistance, matching_index  # noqa: E999
 
-            similarity = fuzz.token_sort_ratio(name, possible_name)
-            if similarity > max_similarity:
-                max_similarity = similarity
+            distance = self.distance(name, possible_name)
+            if mindistance is None or distance < mindistance:
+                mindistance = distance
                 matching_index = i
 
         for i, possible_name in enumerate(possible_names):
@@ -51,6 +51,6 @@ class Phonetics:
                 check_name(name, transformed_possible_name)
                 if alternative_name:
                     check_name(alternative_name, transformed_possible_name)
-        if max_similarity < threshold:
+        if mindistance is None or mindistance > threshold:
             return None
         return matching_index

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/tests/fixtures/adminlevel.yaml

@@ -434,7 +434,6 @@ admin_info:
   - {pcode: XYZ123456, name: Random, iso3: XYZ}
 
 countries_fuzzy_try:
-  - AFG
   - NER
   - NGA
   - UKR

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/tests/fixtures/adminlevelparent.yaml

@@ -16,11 +16,10 @@ admin_name_mappings:
   "AF05|MyMapping3": "AF0501"
 
 admin_name_replacements:
-  "wx": "an"
+  " city": ""
 
 alt1_admin_name_replacements:
-  "COD|wx": "an"
+  "COD| city": ""
 
 alt2_admin_name_replacements:
-  "CD20|wx": "an"
-  "CD31|wx": "en"
+  "CD20| city": ""

{hdx_python_country-3.7.3 → hdx_python_country-3.7.5}/tests/hdx/location/test_adminlevel.py

@@ -196,21 +196,12 @@ class TestAdminLevel:
             "YE30",
             False,
         )
-        assert adminone.get_pcode("YEM", "Ad Jauf", logname="test") == (
-            "YE16",
-            False,
-        )
-        assert adminone.get_pcode("AFG", "Sar-E-Pul", logname="test") == (
-            "AF22",
-            False,
-        )
         assert adminone.get_pcode("SOM", "Bay", logname="test") == (
             "SO24",
             True,
         )
         output = adminone.output_matches()
         assert output == [
-            "test - YEM: Matching (fuzzy) Ad Jauf to Al Jawf on map",
             "test - YEM: Matching (fuzzy) Al Dali to Ad Dali on map",
             "test - YEM: Matching (fuzzy) Al Dhale'e / الضالع to Ad Dali on map",
             "test - YEM: Matching (fuzzy) Al_Dhale'a to Ad Dali on map",
@@ -295,19 +286,21 @@ class TestAdminLevel:
         ) == (None, False)
 
         output = admintwo.output_admin_name_replacements()
-        assert output == ["wx: an"]
+        assert output == [" city: "]
         assert admintwo.get_pcode(
-            "COD", "City of Mbwxza-Nanga", logname="test"
+            "COD", "Mbanza-Ngungu city", logname="test"
         ) == ("CD2013", False)
         assert admintwo.get_pcode(
-            "COD", "City of Mbwxza-Nanga", parent="CD20", logname="test"
+            "COD", "Mbanza-Ngungu city", parent="CD20", logname="test"
         ) == ("CD2013", False)
         assert admintwo.get_pcode(
-            "COD", "City of Mbwxza-Nanga", parent="CD19", logname="test"
+            "COD", "Mbanza-Ngungu city", parent="CD19", logname="test"
         ) == (None, False)
-        assert admintwo.get_pcode(
-            "MWI", "City of Blwxtyre", logname="test"
-        ) == (
+        assert admintwo.get_pcode("COD", "Kenge city", logname="test") == (
+            "CD3102",
+            False,
+        )
+        assert admintwo.get_pcode("MWI", "Blantyre city", logname="test") == (
             "MW305",
             False,
         )
@@ -316,66 +309,58 @@ class TestAdminLevel:
             "alt1_admin_name_replacements"
         ]
         output = admintwo.output_admin_name_replacements()
-        assert output == ["COD|wx: an"]
+        assert output == ["COD| city: "]
         assert admintwo.get_pcode(
-            "COD", "City of Mbwxza-Nanga", logname="test"
+            "COD", "Mbanza-Ngungu city", logname="test"
         ) == ("CD2013", False)
         assert admintwo.get_pcode(
-            "COD", "City of Mbwxza-Nanga", parent="CD20", logname="test"
+            "COD", "Mbanza-Ngungu city", parent="CD20", logname="test"
         ) == ("CD2013", False)
         assert admintwo.get_pcode(
-            "COD", "City of Mbwxza-Nanga", parent="CD19", logname="test"
+            "COD", "Mbanza-Ngungu city", parent="CD19", logname="test"
         ) == (None, False)
-        assert admintwo.get_pcode("COD", "Kenge City", logname="test") == (
+        assert admintwo.get_pcode("COD", "Kenge city", logname="test") == (
             "CD3102",
             False,
         )
-        assert admintwo.get_pcode("COD", "Kwxge City", logname="test") == (
-            None,
-            False,
-        )
         assert admintwo.get_pcode(
-            "COD", "Kwxge City", parent="CD31", logname="test"
-        ) == (None, False)
-        assert admintwo.get_pcode(
-            "MWI", "City of Blwxtyre", logname="test"
-        ) == (
+            "COD", "Kenge city", parent="CD31", logname="test"
+        ) == ("CD3102", False)
+        assert admintwo.get_pcode("MWI", "Blantyre city", logname="test") == (
             None,
             False,
         )
         assert admintwo.get_pcode(
-            "MWI", "City of Blwxtyre", parent="MW3", logname="test"
+            "MWI", "Blantyre city", parent="MW3", logname="test"
         ) == (None, False)
 
         admintwo.admin_name_replacements = config_parent[
             "alt2_admin_name_replacements"
         ]
         output = admintwo.output_admin_name_replacements()
-        assert output == ["CD20|wx: an", "CD31|wx: en"]
+        assert output == ["CD20| city: "]
         assert admintwo.get_pcode(
-            "COD", "City of Mbwxza-Nanga", logname="test"
+            "COD", "Mbanza-Ngungu city", logname="test"
         ) == (None, False)
         assert admintwo.get_pcode(
-            "COD", "City of Mbwxza-Nanga", parent="CD20", logname="test"
+            "COD", "Mbanza-Ngungu city", parent="CD20", logname="test"
         ) == ("CD2013", False)
         assert admintwo.get_pcode(
-            "COD", "City of Mbwxza-Nanga", parent="CD19", logname="test"
+            "COD", "Mbanza-Ngungu city", parent="CD19", logname="test"
         ) == (None, False)
-        assert admintwo.get_pcode("COD", "Kwxge City", logname="test") == (
+        assert admintwo.get_pcode("COD", "Kenge city", logname="test") == (
             None,
             False,
         )
         assert admintwo.get_pcode(
-            "COD", "Kwxge City", parent="CD31", logname="test"
-        ) == ("CD3102", False)
-        assert admintwo.get_pcode(
-            "MWI", "City of Blwxtyre", logname="test"
-        ) == (
+            "COD", "Kenge city", parent="CD31", logname="test"
+        ) == (None, False)
+        assert admintwo.get_pcode("MWI", "Blantyre city", logname="test") == (
             None,
             False,
         )
         assert admintwo.get_pcode(
-            "MWI", "City of Blwxtyre", parent="MW3", logname="test"
+            "MWI", "Blantyre city", parent="MW3", logname="test"
         ) == (None, False)
 
     def test_adminlevel_with_url(self, config, url):

hdx_python_country-3.7.3/src/hdx/location/names.py

@@ -1,31 +0,0 @@
-import re
-import unicodedata
-
-from unidecode import unidecode
-
-non_ascii = r"([^\x00-\x7f])+"
-
-
-def clean_name(name: str) -> str:
-    """
-    Replace accented characters with non-accented ones in given name. Also
-    strip whitespace from start and end and make lwoercase.
-
-    Args:
-        name (str): Name to change
-
-    Returns:
-        str: Name without accented characters
-    """
-    #
-    clean_name = "".join(
-        c
-        for c in unicodedata.normalize("NFD", name)
-        if unicodedata.category(c) != "Mn"
-    )
-    # Remove all non-ASCII characters
-    clean_name = re.sub(non_ascii, " ", clean_name)
-    clean_name = clean_name.replace("'", "")
-    clean_name = re.sub(r"[\W_]", " ", clean_name)
-    clean_name = unidecode(clean_name)
-    return clean_name.strip().lower()