hdx-python-country 3.8.0__tar.gz → 3.8.2__tar.gz

{hdx_python_country-3.8.0 → hdx_python_country-3.8.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: hdx-python-country
-Version: 3.8.0
+Version: 3.8.2
 Summary: HDX Python country code and exchange rate (fx) utilities
 Project-URL: Homepage, https://github.com/OCHA-DAP/hdx-python-country
 Author: Michael Rans

{hdx_python_country-3.8.0 → hdx_python_country-3.8.2}/requirements.txt

@@ -14,13 +14,13 @@ cfgv==3.4.0
     # via pre-commit
 chardet==5.2.0
     # via frictionless
-charset-normalizer==3.3.2
+charset-normalizer==3.4.0
     # via requests
 click==8.1.7
     # via typer
-coverage==7.6.1
+coverage==7.6.3
     # via pytest-cov
-distlib==0.3.8
+distlib==0.3.9
     # via virtualenv
 et-xmlfile==1.1.0
     # via openpyxl
@@ -40,19 +40,19 @@ ijson==3.3.0
     # via hdx-python-utilities
 iniconfig==2.0.0
     # via pytest
-isodate==0.6.1
+isodate==0.7.2
     # via frictionless
 jinja2==3.1.4
     # via frictionless
 jsonlines==4.0.0
     # via hdx-python-utilities
-jsonpath-ng==1.6.1
+jsonpath-ng==1.7.0
     # via libhxl
 jsonschema==4.23.0
     # via
     #   frictionless
     #   tableschema-to-template
-jsonschema-specifications==2023.12.1
+jsonschema-specifications==2024.10.1
     # via jsonschema
 libhxl==5.2.1
     # via hdx-python-country (pyproject.toml)
@@ -62,7 +62,7 @@ markdown-it-py==3.0.0
     # via rich
 marko==2.1.2
     # via frictionless
-markupsafe==2.1.5
+markupsafe==3.0.1
     # via jinja2
 mdurl==0.1.2
     # via markdown-it-py
@@ -82,7 +82,7 @@ ply==3.11
     # via
     #   jsonpath-ng
     #   libhxl
-pre-commit==4.0.0
+pre-commit==4.0.1
     # via hdx-python-country (pyproject.toml)
 pydantic==2.9.2
     # via frictionless
@@ -142,9 +142,7 @@ shellingham==1.5.4
 simpleeval==1.0.0
     # via frictionless
 six==1.16.0
-    # via
-    #   isodate
-    #   python-dateutil
+    # via python-dateutil
 stringcase==1.2.0
     # via frictionless
 structlog==24.4.0

{hdx_python_country-3.8.0 → hdx_python_country-3.8.2}/src/hdx/location/_version.py

@@ -12,5 +12,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '3.8.0'
-__version_tuple__ = version_tuple = (3, 8, 0)
+__version__ = version = '3.8.2'
+__version_tuple__ = version_tuple = (3, 8, 2)

{hdx_python_country-3.8.0 → hdx_python_country-3.8.2}/src/hdx/location/adminlevel.py

@@ -50,9 +50,10 @@ class AdminLevel:
 
     pcode_regex = re.compile(r"^([a-zA-Z]{2,3})(\d+)$")
     _admin_url_default = "https://data.humdata.org/dataset/cb963915-d7d1-4ffa-90dc-31277e24406f/resource/f65bc260-4d8b-416f-ac07-f2433b4d5142/download/global_pcodes_adm_1_2.csv"
-    _admin_url = _admin_url_default
+    admin_url = _admin_url_default
+    admin_all_pcodes_url = "https://data.humdata.org/dataset/cb963915-d7d1-4ffa-90dc-31277e24406f/resource/793e66fe-4cdb-4076-b037-fb8c053239e2/download/global_pcodes.csv"
     _formats_url_default = "https://data.humdata.org/dataset/cb963915-d7d1-4ffa-90dc-31277e24406f/resource/f1161807-dab4-4331-b7b0-4e5dac56e0e4/download/global_pcode_lengths.csv"
-    _formats_url = _formats_url_default
+    formats_url = _formats_url_default
 
     def __init__(
         self,
@@ -113,11 +114,11 @@ class AdminLevel:
         """
         if admin_url is None:
             admin_url = cls._admin_url_default
-        cls._admin_url = admin_url
+        cls.admin_url = admin_url
 
     @staticmethod
     def get_libhxl_dataset(
-        url: str = _admin_url, retriever: Optional[Retrieve] = None
+        url: str = admin_url, retriever: Optional[Retrieve] = None
     ) -> hxl.Dataset:
         """
         Get libhxl Dataset object given a URL which defaults to global p-codes
@@ -262,7 +263,7 @@ class AdminLevel:
 
     def setup_from_url(
         self,
-        admin_url: str = _admin_url,
+        admin_url: str = admin_url,
         countryiso3s: Optional[ListTuple[str]] = None,
     ) -> None:
         """
@@ -278,18 +279,19 @@ class AdminLevel:
         admin_info = self.get_libhxl_dataset(admin_url, self.retriever)
         self.setup_from_libhxl_dataset(admin_info, countryiso3s)
 
-    def load_pcode_formats(self, formats_url: str = _formats_url) -> None:
+    def load_pcode_formats_from_libhxl_dataset(
+        self, libhxl_dataset: hxl.Dataset
+    ) -> None:
         """
-        Load p-code formats from a URL. Defaults to global p-codes dataset on HDX.
+        Load p-code formats from a libhxl Dataset object.
 
         Args:
-            formats_url (str): URL from which to load data. Defaults to global p-codes dataset.
+            libhxl_dataset (hxl.Dataset): Dataset object from libhxl library
 
         Returns:
             None
         """
-        formats_info = self.get_libhxl_dataset(formats_url, self.retriever)
-        for row in formats_info:
+        for row in libhxl_dataset:
             pcode_format = [int(row.get("#country+len"))]
             for admin_no in range(1, 4):
                 length = row.get(f"#adm{admin_no}+len")
@@ -303,6 +305,19 @@ class AdminLevel:
             for x in re.finditer("0", pcode):
                 dict_of_sets_add(self.zeroes, countryiso3, x.start())
 
+    def load_pcode_formats(self, formats_url: str = formats_url) -> None:
+        """
+        Load p-code formats from a URL. Defaults to global p-codes dataset on HDX.
+
+        Args:
+            formats_url (str): URL from which to load data. Defaults to global p-codes dataset.
+
+        Returns:
+            None
+        """
+        formats_info = self.get_libhxl_dataset(formats_url, self.retriever)
+        self.load_pcode_formats_from_libhxl_dataset(formats_info)
+
     def set_parent_admins(self, parent_admins: List[List]) -> None:
         """
         Set parent admins
@@ -587,6 +602,38 @@ class AdminLevel:
                 continue
         return relevant_name_replacements
 
+    def get_admin_fuzzy_dont(
+        self, countryiso3: str, parent: Optional[str]
+    ) -> List[str]:
+        """Get relevant admin names that should not be fuzzy matched from
+        admin fuzzy dont which is a list of strings. These can be global
+        or they can be restricted by country or parent. Keys take the form
+        "DONT_MATCH", "AFG|DONT_MATCH", or "AF01|DONT_MATCH".
+
+        Args:
+            countryiso3 (str): ISO3 country code
+            parent (Optional[str]): Parent admin code
+
+        Returns:
+            List[str]: Relevant admin names that should not be fuzzy matched
+        """
+        relevant_admin_fuzzy_dont = []
+        for value in self.admin_fuzzy_dont:
+            if "|" not in value:
+                if value not in relevant_admin_fuzzy_dont:
+                    relevant_admin_fuzzy_dont.append(value)
+                continue
+            prefix, name = value.split("|")
+            if parent:
+                if prefix == parent:
+                    if name not in relevant_admin_fuzzy_dont:
+                        relevant_admin_fuzzy_dont.append(name)
+            if prefix == countryiso3:
+                if name not in relevant_admin_fuzzy_dont:
+                    relevant_admin_fuzzy_dont.append(name)
+                continue
+        return relevant_admin_fuzzy_dont
+
     def fuzzy_pcode(
         self,
         countryiso3: str,
@@ -643,7 +690,9 @@ class AdminLevel:
         pcode = name_to_pcode.get(
             normalised_name, name_to_pcode.get(alt_normalised_name)
         )
-        if not pcode and name.lower() in self.admin_fuzzy_dont:
+        if not pcode and name.lower() in self.get_admin_fuzzy_dont(
+            countryiso3, parent
+        ):
             if logname:
                 self.ignored.add((logname, countryiso3, name))
             return None

{hdx_python_country-3.8.0 → hdx_python_country-3.8.2}/tests/fixtures/adminlevel.yaml

@@ -516,7 +516,7 @@ admin_name_replacements:
   " oblast": ""
 
 admin_fuzzy_dont:
-  - "nord"
+  - "YEM|nord"
   - "north"
   - "sud"
   - "south"

{hdx_python_country-3.8.0 → hdx_python_country-3.8.2}/tests/hdx/location/test_adminlevel.py

@@ -381,9 +381,9 @@ class TestAdminLevel:
             adminone.setup_from_url("fake_url")
         adminone = AdminLevel(config)
         AdminLevel.set_default_admin_url()
-        assert AdminLevel._admin_url == AdminLevel._admin_url_default
+        assert AdminLevel.admin_url == AdminLevel._admin_url_default
         AdminLevel.set_default_admin_url(url)
-        assert AdminLevel._admin_url == url
+        assert AdminLevel.admin_url == url
         adminone.setup_from_url(countryiso3s=("YEM",))
         assert len(adminone.get_pcode_list()) == 22
 
@@ -421,7 +421,7 @@ class TestAdminLevel:
         adminone = AdminLevel(config)
         adminone.setup_from_url()
         assert adminone.get_admin_level("YEM") == 1
-        assert len(adminone.get_pcode_list()) == 2510
+        assert len(adminone.get_pcode_list()) == 2526
         assert adminone.get_pcode_length("YEM") == 4
         assert adminone.get_pcode("YEM", "YE30", logname="test") == (
             "YE30",