hdx-python-country 3.7.5__tar.gz → 3.7.7__tar.gz

{hdx_python_country-3.7.5 → hdx_python_country-3.7.7}/.config/pre-commit-config.yaml

@@ -8,7 +8,7 @@ repos:
       - id: end-of-file-fixer
       - id: check-ast
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.4.8
+    rev: v0.5.1
     hooks:
       # Run the linter.
       - id: ruff
@@ -17,7 +17,7 @@ repos:
       - id: ruff-format
         args: [--config, .config/ruff.toml]
   - repo: https://github.com/astral-sh/uv-pre-commit
-    rev: 0.2.10
+    rev: 0.2.13
     hooks:
       # Run the pip compile
       - id: pip-compile

{hdx_python_country-3.7.5 → hdx_python_country-3.7.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: hdx-python-country
-Version: 3.7.5
+Version: 3.7.7
 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-3.7.5 → hdx_python_country-3.7.7}/requirements.txt

@@ -18,7 +18,7 @@ charset-normalizer==3.3.2
     # via requests
 click==8.1.7
     # via typer
-coverage==7.5.4
+coverage==7.6.0
     # via pytest-cov
 distlib==0.3.8
     # via virtualenv
@@ -30,9 +30,9 @@ frictionless==5.17.0
     # via hdx-python-utilities
 hdx-python-utilities==3.7.2
     # via hdx-python-country (pyproject.toml)
-humanize==4.9.0
+humanize==4.10.0
     # via frictionless
-identify==2.5.36
+identify==2.6.0
     # via pre-commit
 idna==3.7
     # via requests
@@ -48,7 +48,7 @@ jsonlines==4.0.0
     # via hdx-python-utilities
 jsonpath-ng==1.6.1
     # via libhxl
-jsonschema==4.22.0
+jsonschema==4.23.0
     # via
     #   frictionless
     #   tableschema-to-template
@@ -92,7 +92,7 @@ pygments==2.18.0
     # via rich
 pyphonetics==0.5.3
     # via hdx-python-country (pyproject.toml)
-pytest==8.2.2
+pytest==8.3.1
     # via
     #   hdx-python-country (pyproject.toml)
     #   pytest-cov
@@ -129,7 +129,7 @@ rfc3986==2.0.0
     # via frictionless
 rich==13.7.1
     # via typer
-rpds-py==0.18.1
+rpds-py==0.19.0
     # via
     #   jsonschema
     #   referencing
@@ -147,7 +147,7 @@ six==1.16.0
     #   python-dateutil
 stringcase==1.2.0
     # via frictionless
-structlog==24.2.0
+structlog==24.4.0
     # via libhxl
 tableschema-to-template==0.0.13
     # via hdx-python-utilities
@@ -171,7 +171,7 @@ urllib3==2.2.2
     # via
     #   libhxl
     #   requests
-validators==0.30.0
+validators==0.33.0
     # via frictionless
 virtualenv==20.26.3
     # via pre-commit

{hdx_python_country-3.7.5 → hdx_python_country-3.7.7}/src/hdx/location/_version.py

@@ -12,5 +12,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '3.7.5'
-__version_tuple__ = version_tuple = (3, 7, 5)
+__version__ = version = '3.7.7'
+__version_tuple__ = version_tuple = (3, 7, 7)

{hdx_python_country-3.7.5 → hdx_python_country-3.7.7}/src/hdx/location/adminlevel.py

@@ -125,7 +125,7 @@ class AdminLevel:
         self,
         countryiso3: str,
         pcode: str,
-        adm_name: str,
+        adm_name: Optional[str],
         parent: Optional[str],
     ):
         """
@@ -134,7 +134,7 @@ class AdminLevel:
         Args:
             countryiso3 (str): Country
             pcode (str): P-code
-            adm_name (str): Administrative name
+            adm_name (Optional[str]): Administrative name (which can be None)
             parent (Optional[str]): Parent p-code
 
         Returns:
@@ -142,20 +142,27 @@ class AdminLevel:
         """
         self.pcode_lengths[countryiso3] = len(pcode)
         self.pcodes.append(pcode)
+        if adm_name is None:
+            adm_name = ""
         self.pcode_to_name[pcode] = adm_name
+        self.pcode_to_iso3[pcode] = countryiso3
+        if not adm_name:
+            logger.error(
+                f"Admin name is blank for pcode {pcode} of {countryiso3}!"
+            )
+            return
 
+        adm_name = normalise(adm_name)
         name_to_pcode = self.name_to_pcode.get(countryiso3, {})
-        name_to_pcode[normalise(adm_name)] = pcode
+        name_to_pcode[adm_name] = pcode
         self.name_to_pcode[countryiso3] = name_to_pcode
-        self.pcode_to_iso3[pcode] = countryiso3
-        self.pcode_to_iso3[pcode] = countryiso3
 
         if self.use_parent:
             name_parent_to_pcode = self.name_parent_to_pcode.get(
                 countryiso3, {}
             )
             name_to_pcode = name_parent_to_pcode.get(parent, {})
-            name_to_pcode[normalise(adm_name)] = pcode
+            name_to_pcode[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
@@ -554,6 +561,7 @@ class AdminLevel:
         self,
         countryiso3: str,
         name: str,
+        normalised_name: str,
         **kwargs: Any,
     ) -> Optional[str]:
         """Fuzzy match name to pcode
@@ -561,6 +569,7 @@ class AdminLevel:
         Args:
             countryiso3 (str): ISO3 country code
             name (str): Name to match
+            normalised_name (str): Normalised name
             **kwargs:
             parent (Optional[str]): Parent admin code
             logname (str): Log using this identifying name. Defaults to not logging.
@@ -597,13 +606,12 @@ class AdminLevel:
                 if logname:
                     self.errors.add((logname, countryiso3, parent))
                 return None
-        adm_name_lookup = normalise(name)
-        adm_name_lookup2 = multiple_replace(
-            adm_name_lookup,
+        alt_normalised_name = multiple_replace(
+            normalised_name,
             self.get_admin_name_replacements(countryiso3, parent),
         )
         pcode = name_to_pcode.get(
-            adm_name_lookup, name_to_pcode.get(adm_name_lookup2)
+            normalised_name, name_to_pcode.get(alt_normalised_name)
         )
         if not pcode and name.lower() in self.admin_fuzzy_dont:
             if logname:
@@ -611,7 +619,7 @@ class AdminLevel:
             return None
         if not pcode:
             for map_name in name_to_pcode:
-                if adm_name_lookup in map_name:
+                if normalised_name in map_name:
                     pcode = name_to_pcode[map_name]
                     if logname:
                         self.matches.add(
@@ -625,7 +633,7 @@ class AdminLevel:
                         )
                     break
             for map_name in name_to_pcode:
-                if adm_name_lookup2 in map_name:
+                if alt_normalised_name in map_name:
                     pcode = name_to_pcode[map_name]
                     if logname:
                         self.matches.add(
@@ -659,8 +667,8 @@ class AdminLevel:
 
             matching_index = self.phonetics.match(
                 map_names,
-                adm_name_lookup,
-                alternative_name=adm_name_lookup2,
+                normalised_name,
+                alternative_name=alt_normalised_name,
                 transform_possible_names=[al_transform_1, al_transform_2],
             )
 
@@ -754,6 +762,7 @@ class AdminLevel:
             )
             return pcode, True
         else:
+            normalised_name = normalise(name)
             if parent:
                 name_parent_to_pcode = self.name_parent_to_pcode.get(
                     countryiso3
@@ -761,18 +770,20 @@ class AdminLevel:
                 if name_parent_to_pcode:
                     name_to_pcode = name_parent_to_pcode.get(parent)
                     if name_to_pcode is not None:
-                        pcode = name_to_pcode.get(name.lower())
+                        pcode = name_to_pcode.get(normalised_name)
                         if pcode:
                             return pcode, True
             else:
                 name_to_pcode = self.name_to_pcode.get(countryiso3)
                 if name_to_pcode is not None:
-                    pcode = name_to_pcode.get(name.lower())
+                    pcode = name_to_pcode.get(normalised_name)
                     if pcode:
                         return pcode, True
-            if not fuzzy_match or len(name) < fuzzy_length:
+            if not fuzzy_match or len(normalised_name) < fuzzy_length:
                 return None, True
-            pcode = self.fuzzy_pcode(countryiso3, name, **kwargs)
+            pcode = self.fuzzy_pcode(
+                countryiso3, name, normalised_name, **kwargs
+            )
             return pcode, False
 
     def output_matches(self) -> List[str]:

hdx_python_country-3.7.7/tests/fixtures/global_pcode_lengths.csv

@@ -0,0 +1,158 @@
+Location,Country Length,Admin 1 Length,Admin 2 Length,Admin 3 Length,Admin 4 Length,Admin 5 Length
+#country+code,#country+len,#adm1+len,#adm2+len,#adm3+len,#adm4+len,#adm5+len
+AFG,2,2,2,,,
+AGO,2,2,3,3,,
+ALB,2,2,2,2,,
+ARE,2,2,,,,
+ARG,2,3,3,,,
+ARM,2,2,1,3,,
+ATG,2,2,,,,
+AZE,2,8,,,,
+BDI,3,3,3,,,
+BEN,2,2,2,,,
+BES,3,1,,,,
+BFA,2,2,2,2,,
+BGD,2,2,2,2,2,
+BGR,2,3,3,,,
+BLR,2,3,3,,,
+BLZ,2,2,,,,
+BMU,2,2,1,,,
+BOL,2,2,2,2,,
+BRA,2,2,5,,,
+BRB,2,2,,,,
+BTN,2,3,2,,,
+BWA,2,2,2,2,,
+CAF,2,2,1,1,2,
+CHL,2,2,1,2,,
+CHN,2,3,3,,,
+CIV,2,2,2,2,,
+CMR,2,3,3,3,,
+COD,2,2,2,,,
+COG,2,2,2,,,
+COL,2,2,3,,,
+COM,2,1,1,1,,
+CPV,2,2,2,,,
+CRI,2,1,2,2,,
+CUB,2,2,2,,,
+CUW,2,2,,,,
+CYM,2,2,,,,
+DJI,2,2,2,,,
+DMA,2,2,,,,
+DOM,2,2,2,2,2,
+DZA,2,3,3,,,
+ECU,2,2,2,2,,
+EGY,2,2,2,,,
+ERI,2,1,2,,,
+ESH,2,2,,,,
+ETH,2,2,2,2,,
+FJI,2,1,2,2,,
+FSM,2,1,2,,,
+GAB,3,3,3,,,
+GEO,2,2,2,,,
+GHA,2,2,2,,,
+GIN,2,3,3,2,,
+GLP,2,2,2,,,
+GMB,2,2,2,2,,
+GNB,2,2,2,,,
+GNQ,2,3,3,,,
+GRD,2,2,,,,
+GTM,2,2,2,,,
+GUF,2,1,3,,,
+GUY,2,2,2,,,
+HND,2,2,2,,,
+HTI,2,2,2,3,,
+HUN,2,3,3,,,
+IDN,2,2,2,3,,
+IRN,2,3,3,,,
+IRQ,2,3,3,3,,
+JAM,2,2,3,,,
+KAZ,3,3,3,,,
+KEN,2,3,3,,,
+KGZ,2,11,0,0,,
+KHM,2,2,2,2,,
+KIR,2,1,2,,,
+KNA,2,2,,,,
+KWT,2,2,,,,
+LAO,2,2,2,,,
+LBN,2,1,1,1,,
+LBR,2,2,2,,,
+LBY,2,2,2,,,
+LCA,2,2,9,,,
+LKA,2,1,1,2,3,
+LSO,2,1,2,,,
+MAR,2,3,3,7,1,
+MDA,2,3,,,,
+MDG,2,2,7|3,-1|3,3,
+MDV,2,3,3,3,,
+MEX,2,2,3,,,
+MHL,2,2,2,,,
+MLI,2,2,2,2,,
+MMR,3,3,4,-1,3,3
+MNG,2,2,2,,,
+MOZ,2,2,2,2,,
+MRT,2,2,1,2,,
+MSR,2,2,,,,
+MTQ,2,2,2,,,
+MUS,2,2,,,,
+MWI,2,1,2,2,,
+MYS,2,2,2,,,
+NAM,2,2,2,,,
+NER,2,3,3,3,,
+NGA,2,3,3,3,,
+NIC,2,2,2,,,
+NPL,2,2,2,3,,
+OMN,2,2,2,,,
+PAK,2,1,2,2,,
+PAN,2,2,2,2,,
+PER,2,2,2,2,,
+PHL,2,2,3,2,3,
+PNG,2,2,2,2,,
+POL,2,3,3,,,
+PRI,2,2,,,,
+PRK,2,2,2,,,
+PRY,2,2,2,,,
+PSE,2,2,2,,,
+QAT,3,3,3,,,
+ROU,2,3,3,,,
+RUS,2,3,3,,,
+RWA,2,1,1,2,2,
+SAU,2,2,,,,
+SDN,2,2,3,,,
+SEN,2,2,2,2,,
+SLB,2,2,4,4,,
+SLE,2,2,2,2,2,
+SLV,2,2,2,,,
+SOM,2,2,2,,,
+SSD,2,2,2,2,,
+STP,2,2,2,,,
+SUR,2,2,2,,,
+SVK,2,3,3,,,
+SWZ,2,1,2,,,
+SXM,2,1,,,,
+SYC,2,1,1,4,,
+SYR,2,2,2,2,,
+TCA,2,1,2,,,
+TCD,2,2,2,2,,
+TGO,2,2,2,2,,
+THA,2,2,2,2,,
+TJK,0,7,0,,,
+TLS,2,2,2,2,,
+TON,2,1,1,2,,
+TTO,2,2,,,,
+TUN,2,1,1,2,2,
+TUR,3,3,3,3,3,
+TZA,2,2,2,3|4,,
+UGA,2,1,3,2,2,
+UKR,2,2,2,3,3,
+URY,2,2,3,,,
+UZB,2,2,3,,,
+VCT,2,1,2,,,
+VEN,2,2,2,2,,
+VGB,2,2,,,,
+VIR,2,3,5,,,
+VNM,2,3,2,,,
+VUT,2,2,3,,,
+YEM,2,2,2,2,,
+ZAF,2,1,2,1,3,
+ZMB,2,3,3,3,3,
+ZWE,2,2,2,2,,