build universal

.github/workflows/wheels.yml

@@ -25,6 +25,7 @@ jobs:
         run: python -m cibuildwheel --output-dir wheelhouse
         env:
           CIBW_SKIP: cp36-* cp37-* *musllinux*
+          CIBW_ARCHS_MACOS: "x86_64 universal2"
           CIBW_BEFORE_ALL_LINUX: yum install -y autoconf automake curl libtool sudo
           CIBW_BEFORE_ALL_MACOS: brew install autoconf automake curl libtool pkg-config
 

install_libpostal.sh

@@ -1,7 +1,6 @@
 #!/bin/bash
 
 OS=$(uname -s)
-ARCH=$(uname -m)
 
 case $OS in
     Linux)
@@ -22,16 +21,6 @@ case $OS in
         ;;
 esac
 
-if [ "$OS" = "Linux" ]; then
-    sudo apt-get install clang curl autoconf automake libtool pkg-config
-elif [ "$OS" = "Darwin" ]; then
-    if [ "$ARCH" == "arm64" ]; then
-        arch -arm64 brew install curl autoconf automake libtool pkg-config
-    else
-        brew install curl autoconf automake libtool pkg-config
-    fi
-fi
-
 cd $1
 git clone https://github.com/openvenues/libpostal
 cd libpostal

libpypostal/parser.py

@@ -2,8 +2,6 @@
 from enum import Enum
 from typing import Dict, List, Optional, Tuple
 
-from libpypostal import _parser  # type: ignore # pylint: disable=no-name-in-module
-
 
 class LibpostalAddressComponent(str, Enum):
     """Libpostal address component."""
@@ -30,6 +28,16 @@ class LibpostalAddressComponent(str, Enum):
     WORLD_REGION = "world_region"
 
 
+def _parse_address(
+    address: str, language: str, country_code: str
+) -> List[Tuple[str, str]]:
+    from libpypostal import _parser  # type: ignore # pylint: disable=no-name-in-module,import-outside-toplevel
+
+    return _parser.parse_address(  # pylint: disable=c-extension-no-member
+        address, language=language, country=country_code
+    )
+
+
 def parse_address(
     address: str, language: Optional[str] = None, country_code: Optional[str] = None
 ) -> Dict[str, List[str]]:
@@ -46,10 +54,8 @@ def parse_address(
         possibility of multiple matches. Address components not found in the input are
         set to empty lists.
     """
-    address_component_tuples: List[
-        Tuple[str, str]
-    ] = _parser.parse_address(  # pylint: disable=c-extension-no-member
-        address, language=language, country=country_code
+    address_component_tuples = _parse_address(
+        address, language=language, country_code=country_code
     )
 
     parsed_address_components: Dict[str, List[str]] = {}

tests/test_parser.py

@@ -0,0 +1,93 @@
+# pylint: disable=missing-module-docstring,missing-function-docstring,
+from unittest.mock import MagicMock, patch
+
+from libpypostal import parser
+
+
+@patch("libpypostal._parse_address")
+def test_parse_address_single_match_each_component(
+    mock__parser_parse_address: MagicMock,
+) -> None:
+    test_address = "123 Bayes Way, Beverly Hills, CA 90210 US"
+    mock__parser_parse_address.return_value = [
+        ("123", "house_number"),
+        ("bayes way", "road"),
+        ("beverly hills", "city"),
+        ("ca", "state"),
+        ("90210", "postcode"),
+        ("us", "country"),
+    ]
+    expected = {
+        "house_number": ["123"],
+        "road": ["bayes way"],
+        "city": ["beverly hills"],
+        "state": ["ca"],
+        "postcode": ["90210"],
+        "country": ["us"],
+        "category": [],
+        "city_district": [],
+        "country_region": [],
+        "entrance": [],
+        "house": [],
+        "island": [],
+        "level": [],
+        "near": [],
+        "po_box": [],
+        "staircase": [],
+        "state_district": [],
+        "suburb": [],
+        "unit": [],
+        "world_region": [],
+    }
+
+    actual = parser.parse_address(test_address)
+
+    assert actual == expected
+    mock__parser_parse_address.assert_called_once_with(
+        test_address, language=None, country=None
+    )
+
+
+@patch("libpypostal._parse_address")
+def test_parse_address_multiple_matches_for_component(
+    mock__parser_parse_address: MagicMock,
+) -> None:
+    test_address = "123 Bayes Way, Beverly Hills, CA 90210 California US"
+    mock__parser_parse_address.return_value = [
+        ("123", "house_number"),
+        ("bayes way", "road"),
+        ("beverly hills", "city"),
+        ("ca", "state"),
+        ("california", "state"),
+        ("90210", "postcode"),
+        ("us", "country"),
+    ]
+    expected = {
+        "house_number": ["123"],
+        "road": ["bayes way"],
+        "city": ["beverly hills"],
+        "state": ["ca", "california"],
+        "postcode": ["90210"],
+        "country": ["us"],
+        "category": [],
+        "city_district": [],
+        "country_region": [],
+        "entrance": [],
+        "house": [],
+        "island": [],
+        "level": [],
+        "near": [],
+        "po_box": [],
+        "staircase": [],
+        "state_district": [],
+        "suburb": [],
+        "unit": [],
+        "world_region": [],
+    }
+
+    actual = parser.parse_address(test_address)
+
+    assert actual == expected
+    mock__parser_parse_address.assert_called_once_with(
+        test_address, language=None, country=None
+    )