Merge pull request #57 from GNS-Science/feature/56-location-macrons

Feature/56 location macrons

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.8.3
+current_version = 0.8.4
 commit = True
 tag = True
 

CHANGELOG.md

@@ -5,6 +5,14 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.8.4] - 2024-11-20
+
+### Added
+ - function `get_name_with_macrons` to map place names to correct Māori spelling with macrons
+
+### Changed
+ - moved `location` package resources to sub-package and import with `importlib.resources`
+
 ## [0.8.3] - 2024-10-09
 
 ### Changed

nzshm_common/__init__.py

@@ -1,6 +1,6 @@
 __author__ = "GNS Science"
 __email__ = '[email protected]'
-__version__ = '0.8.3'
+__version__ = '0.8.4'
 
 
 # Common classes at the top level for convenience

nzshm_common/location/__init__.py

@@ -25,4 +25,4 @@
 """
 
 from .coded_location import CodedLocation, CodedLocationBin
-from .location import get_location_list, get_location_list_names, get_locations, location_by_id
+from .location import get_location_list, get_location_list_names, get_locations, get_name_with_macrons, location_by_id

nzshm_common/location/location.py

@@ -3,6 +3,7 @@
 """
 
 import csv
+import importlib.resources as resources
 import json
 from collections import namedtuple
 from pathlib import Path
@@ -16,13 +17,15 @@
 # Omitting country for now, focus on NZ
 # https://service.unece.org/trade/locode/nz.htm
 
-locations_filepath = Path(Path(__file__).parent, 'locations.json')
-with open(locations_filepath, 'r') as locations_file:
-    LOCATIONS = json.load(locations_file)
+resource_dir = resources.files('nzshm_common.location.resources')
 
-nz_ids_filepath = Path(Path(__file__).parent, 'nz_ids.json')
-with open(nz_ids_filepath, 'r') as nz_ids_file:
-    NZ_IDS = json.load(nz_ids_file)
+with resources.as_file(resource_dir / 'locations.json') as path:
+    with path.open() as file:
+        LOCATIONS = json.load(file)
+
+with resources.as_file(resource_dir / 'nz_ids.json') as path:
+    with path.open() as file:
+        NZ_IDS = json.load(file)
 
 LOCATIONS_BY_ID: Dict[str, Any] = {location["id"]: location for location in LOCATIONS}
 
@@ -47,12 +50,72 @@
     },
     "ALL": {
         "id": "ALL",
-        "name": "Seismic Risk Working Group NZ code locations",
+        "name": "All locations",
         "locations": list(map(lambda loc: loc["id"], LOCATIONS)),
     },
 }
 
 
+def _get_macron_word_mapping() -> Dict[str, str]:
+    """using the maori_names.csv file as received from LINZ rather than storing the mapping allows
+    us to update without rebuilding the resource"""
+
+    char_map_lower = {
+        'ā': 'a',
+        'ē': 'e',
+        'ī': 'i',
+        'ō': 'o',
+        'ū': 'u',
+    }
+    char_map = {}
+    for k, v in char_map_lower.items():
+        char_map[k] = v
+        char_map[k.upper()] = v.upper()
+    translation_table = str.maketrans(char_map)
+
+    word_mapping = dict()
+    with resources.as_file(resource_dir / 'maori_names.csv') as path:
+        with path.open() as file:
+            reader = csv.reader(file)
+            _ = next(reader)
+            for row in reader:
+                name = row[1]  # second column of LINZ file contains names
+                for word in name.split():  # treat each whole word seperatly
+                    if any([char in char_map for char in word]):  # add to mapping if any characters have macron
+                        word_nomacron = word.translate(translation_table)
+                        word_mapping[word_nomacron] = word
+    return word_mapping
+
+
+WORD_MAPPING = _get_macron_word_mapping()
+
+
+def get_name_with_macrons(name_input: str) -> str:
+    """
+    Corrects the spelling of Māori palce names by adding macrons. Place name spellings from
+    LINZ "Place names of New Zealand".
+    See https://www.linz.govt.nz/products-services/place-names/place-names-new-zealand
+    and https://gazetteer.linz.govt.nz/maori_names.csv
+
+    If the input name is not on the LINZ list, the function will return the input.
+
+    Args:
+        input_name: the name to correct with macrons
+
+    Returns:
+        the place name with the correct Māori spelling
+    """
+
+    words_in = name_input.split()
+    return " ".join([_map_word(word) for word in words_in])
+
+
+def _map_word(word_input):
+    if word_output := WORD_MAPPING.get(word_input):
+        return word_output
+    return word_input
+
+
 def _lat_lon(_id) -> Optional[LatLon]:
     loc = location_by_id(_id)
     if loc:

nzshm_common/location/resources/__init__.py

@@ -0,0 +1 @@
+"""init file needed for python3.9 compatibility"""